Assessing the ability of differential somatic cell count to identify intramammary infections and high somatic cell count at dry-off and in early postpartum

Descriptive Statistics

# About enrolled cows

table1(~as.factor(presence_cm_previous_cow) + as.factor(Lact.SCM) + Lact.group + season.enrollment,data=Doff.by.cow.3 )

	Overall (N=488)
as.factor(presence_cm_previous_cow)
0	328 (67.2%)
1	160 (32.8%)
as.factor(Lact.SCM)
0	284 (58.2%)
1	204 (41.8%)
Lact.group
1	155 (31.8%)
2	187 (38.3%)
3	146 (29.9%)
season.enrollment
W	197 (40.4%)
SP	0 (0%)
SU	49 (10.0%)
A	242 (49.6%)

# ----- Dry-off -----

# Prevalence IMI quarter 

table(Doff.by.quarter.2$Contaminated)

## 
##    0    1 
## 1784  121

prop.table(table(Doff.by.quarter.2$Contaminated))

## 
##          0          1 
## 0.93648294 0.06351706

table(Doff.by.quarter.2$IMI)

## 
##    0    1 
## 1034  750

prop.table(table(Doff.by.quarter.2$IMI))

## 
##         0         1 
## 0.5795964 0.4204036

table(Doff.by.quarter.2$Staph.aureus)

## 
##    0    1 
## 1781    3

prop.table(table(Doff.by.quarter.2$Staph.aureus))

## 
##           0           1 
## 0.998318386 0.001681614

table(Doff.by.quarter.2$CNS)

## 
##    0    1 
## 1327  457

prop.table(table(Doff.by.quarter.2$CNS))

## 
##         0         1 
## 0.7438341 0.2561659

table(Doff.by.quarter.2$SSLO)

## 
##    0    1 
## 1322  462

prop.table(table(Doff.by.quarter.2$SSLO))

## 
##         0         1 
## 0.7410314 0.2589686

table(Doff.by.quarter.2$gram.negative)

## 
##    0    1 
## 1694   90

prop.table(table(Doff.by.quarter.2$gram.negative))

## 
##          0          1 
## 0.94955157 0.05044843

# SCM quarter

table(Doff.by.quarter.2$delaval.SCM)

## 
##    0    1 
## 1529  372

prop.table(table(Doff.by.quarter.2$delaval.SCM))

## 
##         0         1 
## 0.8043135 0.1956865

# Prevalence IMI cow

table(Doff.by.cow.3$Contaminated.cow)

## 
##   0   1 
## 482   6

prop.table(table(Doff.by.cow.3$Contaminated.cow))

## 
##          0          1 
## 0.98770492 0.01229508

table(Doff.by.cow.3$IMI.cow)

## 
##   0   1 
##  98 384

prop.table(table(Doff.by.cow.3$IMI.cow))

## 
##         0         1 
## 0.2033195 0.7966805

table(Doff.by.cow.3$Staph.aureus.cow)

## 
##   0   1 
## 479   3

prop.table(table(Doff.by.cow.3$Staph.aureus.cow))

## 
##           0           1 
## 0.993775934 0.006224066

table(Doff.by.cow.3$CNS.cow)

## 
##   0   1 
## 202 280

prop.table(table(Doff.by.cow.3$CNS.cow))

## 
##         0         1 
## 0.4190871 0.5809129

table(Doff.by.cow.3$SSLO.cow)

## 
##   0   1 
## 192 290

prop.table(table(Doff.by.cow.3$SSLO.cow))

## 
##         0         1 
## 0.3983402 0.6016598

table(Doff.by.cow.3$gram.negative.cow)

## 
##   0   1 
## 404  78

prop.table(table(Doff.by.cow.3$gram.negative.cow))

## 
##         0         1 
## 0.8381743 0.1618257

#SCM cow

table(Doff.by.cow.3$delaval.SCM.CO.COW)

## 
##   0   1 
## 341 145

prop.table(table(Doff.by.cow.3$delaval.SCM.CO.COW))

## 
##         0         1 
## 0.7016461 0.2983539

# ---- C3 -----

# Prevalence IMI quarter 

table(Pc.by.quarter.C3$Contaminated)

## 
##    0    1 
## 1595   90

prop.table(table(Pc.by.quarter.C3$Contaminated))

## 
##          0          1 
## 0.94658754 0.05341246

table(Pc.by.quarter.C3$IMI)

## 
##    0    1 
## 1050  545

prop.table(table(Pc.by.quarter.C3$IMI))

## 
##         0         1 
## 0.6583072 0.3416928

table(Pc.by.quarter.C3$Staph.aureus)

## 
##    0    1 
## 1593    2

prop.table(table(Pc.by.quarter.C3$Staph.aureus))

## 
##           0           1 
## 0.998746082 0.001253918

table(Pc.by.quarter.C3$CNS)

## 
##    0    1 
## 1298  297

prop.table(table(Pc.by.quarter.C3$CNS))

## 
##         0         1 
## 0.8137931 0.1862069

table(Pc.by.quarter.C3$SSLO)

## 
##    0    1 
## 1248  347

prop.table(table(Pc.by.quarter.C3$SSLO))

## 
##         0         1 
## 0.7824451 0.2175549

table(Pc.by.quarter.C3$gram.negative)

## 
##    0    1 
## 1520   75

prop.table(table(Pc.by.quarter.C3$gram.negative))

## 
##          0          1 
## 0.95297806 0.04702194

# SCM quarter

table(Pc.by.quarter.C3$delaval.SCM)

## 
##    0    1 
## 1314  366

prop.table(table(Pc.by.quarter.C3$delaval.SCM))

## 
##         0         1 
## 0.7821429 0.2178571

# Prevalence IMI cow

table(Pc.by.cow.C3$Contaminated.cow)

## 
##   0   1 
## 428   3

prop.table(table(Pc.by.cow.C3$Contaminated.cow))

## 
##           0           1 
## 0.993039443 0.006960557

table(Pc.by.cow.C3$IMI.cow)

## 
##   0   1 
## 115 313

prop.table(table(Pc.by.cow.C3$IMI.cow))

## 
##         0         1 
## 0.2686916 0.7313084

table(Pc.by.cow.C3$Staph.aureus.cow)

## 
##   0   1 
## 426   2

prop.table(table(Pc.by.cow.C3$Staph.aureus.cow))

## 
##           0           1 
## 0.995327103 0.004672897

table(Pc.by.cow.C3$CNS.cow)

## 
##   0   1 
## 219 209

prop.table(table(Pc.by.cow.C3$CNS.cow))

## 
##         0         1 
## 0.5116822 0.4883178

table(Pc.by.cow.C3$SSLO.cow)

## 
##   0   1 
## 196 232

prop.table(table(Pc.by.cow.C3$SSLO.cow))

## 
##         0         1 
## 0.4579439 0.5420561

table(Pc.by.cow.C3$gram.negative.cow)

## 
##   0   1 
## 362  66

prop.table(table(Pc.by.cow.C3$gram.negative.cow))

## 
##         0         1 
## 0.8457944 0.1542056

#SCM cow

table(Pc.by.cow.C3$delaval.SCM.CO.COW)

## 
##   0   1 
## 310 118

prop.table(table(Pc.by.cow.C3$delaval.SCM.CO.COW))

## 
##         0         1 
## 0.7242991 0.2757009

# ---- C5 -----

# Prevalence IMI quarter 

table(Pc.by.quarter.C5$Contaminated)

## 
##    0    1 
## 1563   91

prop.table(table(Pc.by.quarter.C5$Contaminated))

## 
##          0          1 
## 0.94498186 0.05501814

table(Pc.by.quarter.C5$IMI)

## 
##    0    1 
## 1020  543

prop.table(table(Pc.by.quarter.C5$IMI))

## 
##         0         1 
## 0.6525912 0.3474088

table(Pc.by.quarter.C5$Staph.aureus)

## 
##    0    1 
## 1561    2

prop.table(table(Pc.by.quarter.C5$Staph.aureus))

## 
##           0           1 
## 0.998720409 0.001279591

table(Pc.by.quarter.C5$CNS)

## 
##    0    1 
## 1254  309

prop.table(table(Pc.by.quarter.C5$CNS))

## 
##         0         1 
## 0.8023033 0.1976967

table(Pc.by.quarter.C5$SSLO)

## 
##    0    1 
## 1225  338

prop.table(table(Pc.by.quarter.C5$SSLO))

## 
##         0         1 
## 0.7837492 0.2162508

table(Pc.by.quarter.C5$gram.negative)

## 
##    0    1 
## 1491   72

prop.table(table(Pc.by.quarter.C5$gram.negative))

## 
##          0          1 
## 0.95393474 0.04606526

# SCM quarter

table(Pc.by.quarter.C5$delaval.SCM)

## 
##    0    1 
## 1497  145

prop.table(table(Pc.by.quarter.C5$delaval.SCM))

## 
##          0          1 
## 0.91169306 0.08830694

# Prevalence IMI cow

table(Pc.by.cow.C5$Contaminated.cow)

## 
##   0   1 
## 421   3

prop.table(table(Pc.by.cow.C5$Contaminated.cow))

## 
##           0           1 
## 0.992924528 0.007075472

table(Pc.by.cow.C5$IMI.cow)

## 
##   0   1 
## 108 313

prop.table(table(Pc.by.cow.C5$IMI.cow))

## 
##         0         1 
## 0.2565321 0.7434679

table(Pc.by.cow.C5$Staph.aureus.cow)

## 
##   0   1 
## 419   2

prop.table(table(Pc.by.cow.C5$Staph.aureus.cow))

## 
##           0           1 
## 0.995249406 0.004750594

table(Pc.by.cow.C5$CNS.cow)

## 
##   0   1 
## 206 215

prop.table(table(Pc.by.cow.C5$CNS.cow))

## 
##         0         1 
## 0.4893112 0.5106888

table(Pc.by.cow.C5$SSLO.cow)

## 
##   0   1 
## 193 228

prop.table(table(Pc.by.cow.C5$SSLO.cow))

## 
##         0         1 
## 0.4584323 0.5415677

table(Pc.by.cow.C5$gram.negative.cow)

## 
##   0   1 
## 358  63

prop.table(table(Pc.by.cow.C5$gram.negative.cow))

## 
##         0         1 
## 0.8503563 0.1496437

#SCM cow

table(Pc.by.cow.C5$delaval.SCM.CO.COW)

## 
##   0   1 
## 357  63

prop.table(table(Pc.by.cow.C5$delaval.SCM.CO.COW))

## 
##    0    1 
## 0.85 0.15

# DSCC


table1(~log.SCC + term_diag + log.Neu + log.Lym + log.Mac + PERCENT01_N + PERCENT02_N + PERCENT03_N,data=Doff.by.quarter.2)

	Overall (N=1905)
log.SCC
Mean (SD)	4.36 (1.19)
Median [Min, Max]	4.29 [1.10, 8.59]
Missing	4 (0.2%)
term_diag
Negative	1563 (82.0%)
Positive	327 (17.2%)
Missing	15 (0.8%)
log.Neu
Mean (SD)	3.68 (1.26)
Median [Min, Max]	3.50 [0.693, 9.34]
Missing	15 (0.8%)
log.Lym
Mean (SD)	2.66 (1.22)
Median [Min, Max]	2.56 [0, 8.04]
Missing	15 (0.8%)
log.Mac
Mean (SD)	3.09 (1.42)
Median [Min, Max]	2.94 [0, 8.58]
Missing	15 (0.8%)
PERCENT01_N
Mean (SD)	50.3 (12.1)
Median [Min, Max]	50.0 [13.1, 85.7]
Missing	15 (0.8%)
PERCENT02_N
Mean (SD)	19.1 (8.14)
Median [Min, Max]	17.8 [0, 60.0]
Missing	15 (0.8%)
PERCENT03_N
Mean (SD)	30.6 (14.5)
Median [Min, Max]	29.2 [0, 83.5]
Missing	15 (0.8%)

table1(~log.SCC + term_diag + log.Neu + log.Lym + log.Mac + PERCENT01_N + PERCENT02_N + PERCENT03_N,data=Pc.by.quarter.C3)

	Overall (N=1685)
log.SCC
Mean (SD)	4.79 (0.906)
Median [Min, Max]	4.73 [1.79, 8.66]
Missing	5 (0.3%)
term_diag
Negative	1291 (76.6%)
Positive	394 (23.4%)
log.Neu
Mean (SD)	4.14 (0.970)
Median [Min, Max]	4.04 [1.61, 9.43]
Missing	1 (0.1%)
log.Lym
Mean (SD)	3.01 (0.946)
Median [Min, Max]	2.94 [0, 7.22]
Missing	1 (0.1%)
log.Mac
Mean (SD)	3.01 (1.08)
Median [Min, Max]	2.89 [0, 8.99]
Missing	1 (0.1%)
PERCENT01_N
Mean (SD)	58.9 (11.4)
Median [Min, Max]	59.7 [20.2, 87.3]
Missing	1 (0.1%)
PERCENT02_N
Mean (SD)	19.6 (5.98)
Median [Min, Max]	19.2 [2.70, 48.3]
Missing	1 (0.1%)
PERCENT03_N
Mean (SD)	21.4 (11.1)
Median [Min, Max]	19.4 [1.89, 72.9]
Missing	1 (0.1%)

table1(~log.SCC + term_diag + log.Neu + log.Lym + log.Mac + PERCENT01_N + PERCENT02_N + PERCENT03_N,data=Pc.by.quarter.C5)

	Overall (N=1654)
log.SCC
Mean (SD)	4.15 (1.05)
Median [Min, Max]	4.04 [1.39, 8.52]
Missing	12 (0.7%)
term_diag
Negative	1504 (90.9%)
Positive	143 (8.6%)
Missing	7 (0.4%)
log.Neu
Mean (SD)	3.43 (1.11)
Median [Min, Max]	3.26 [1.10, 10.2]
Missing	7 (0.4%)
log.Lym
Mean (SD)	2.43 (1.08)
Median [Min, Max]	2.30 [0, 8.17]
Missing	7 (0.4%)
log.Mac
Mean (SD)	2.35 (1.21)
Median [Min, Max]	2.20 [0, 9.13]
Missing	7 (0.4%)
PERCENT01_N
Mean (SD)	56.2 (10.7)
Median [Min, Max]	56.3 [13.8, 89.6]
Missing	7 (0.4%)
PERCENT02_N
Mean (SD)	21.9 (8.17)
Median [Min, Max]	21.3 [0, 61.5]
Missing	7 (0.4%)
PERCENT03_N
Mean (SD)	21.8 (11.6)
Median [Min, Max]	20.0 [0, 71.4]
Missing	7 (0.4%)

table1(~log.SCC.CO.COW + as.factor(Qscout.CO.COW)  + log.TOTAL.NEU.CO.COW + log.TOTAL.LYM.CO.COW + log.TOTAL.MAC.CO.COW + PERCENT.NEU.CO.COW + PERCENT.LYM.CO.COW + PERCENT.MAC.CO.COW,data=Doff.by.cow.3)

	Overall (N=488)
log.SCC.CO.COW
Mean (SD)	4.68 (1.15)
Median [Min, Max]	4.56 [1.79, 8.05]
Missing	2 (0.4%)
as.factor(Qscout.CO.COW)
0	356 (73.0%)
1	117 (24.0%)
Missing	15 (3.1%)
log.TOTAL.NEU.CO.COW
Mean (SD)	4.03 (1.29)
Median [Min, Max]	3.85 [1.39, 8.67]
Missing	15 (3.1%)
log.TOTAL.LYM.CO.COW
Mean (SD)	2.93 (1.17)
Median [Min, Max]	2.83 [0, 7.08]
Missing	15 (3.1%)
log.TOTAL.MAC.CO.COW
Mean (SD)	3.49 (1.33)
Median [Min, Max]	3.40 [0, 7.29]
Missing	15 (3.1%)
PERCENT.NEU.CO.COW
Mean (SD)	50.7 (12.8)
Median [Min, Max]	51.1 [15.7, 122]
Missing	15 (3.1%)
PERCENT.LYM.CO.COW
Mean (SD)	17.7 (7.39)
Median [Min, Max]	16.9 [3.54, 69.0]
Missing	15 (3.1%)
PERCENT.MAC.CO.COW
Mean (SD)	31.9 (14.3)
Median [Min, Max]	29.9 [3.41, 74.5]
Missing	15 (3.1%)

table1(~log.SCC.CO.COW + as.factor(Qscout.CO.COW)  + log.TOTAL.NEU.CO.COW + log.TOTAL.LYM.CO.COW + log.TOTAL.MAC.CO.COW + PERCENT.NEU.CO.COW + PERCENT.LYM.CO.COW + PERCENT.MAC.CO.COW,data=Pc.by.cow.C3)

	Overall (N=431)
log.SCC.CO.COW
Mean (SD)	5.00 (1.00)
Median [Min, Max]	4.88 [2.83, 8.25]
Missing	3 (0.7%)
as.factor(Qscout.CO.COW)
0	317 (73.5%)
1	113 (26.2%)
Missing	1 (0.2%)
log.TOTAL.NEU.CO.COW
Mean (SD)	4.26 (1.04)
Median [Min, Max]	4.13 [2.20, 8.66]
Missing	1 (0.2%)
log.TOTAL.LYM.CO.COW
Mean (SD)	3.16 (0.983)
Median [Min, Max]	3.07 [0, 6.92]
Missing	1 (0.2%)
log.TOTAL.MAC.CO.COW
Mean (SD)	3.29 (1.12)
Median [Min, Max]	3.18 [0, 8.32]
Missing	1 (0.2%)
PERCENT.NEU.CO.COW
Mean (SD)	57.0 (11.8)
Median [Min, Max]	57.4 [21.2, 87.0]
Missing	1 (0.2%)
PERCENT.LYM.CO.COW
Mean (SD)	19.3 (5.88)
Median [Min, Max]	19.2 [5.08, 51.1]
Missing	1 (0.2%)
PERCENT.MAC.CO.COW
Mean (SD)	23.7 (11.6)
Median [Min, Max]	22.0 [3.51, 73.1]
Missing	1 (0.2%)

table1(~log.SCC.CO.COW + as.factor(Qscout.CO.COW) + log.TOTAL.NEU.CO.COW + log.TOTAL.LYM.CO.COW + log.TOTAL.MAC.CO.COW + PERCENT.NEU.CO.COW + PERCENT.LYM.CO.COW + PERCENT.MAC.CO.COW,data=Pc.by.cow.C5)

	Overall (N=424)
log.SCC.CO.COW
Mean (SD)	4.46 (1.18)
Median [Min, Max]	4.23 [1.39, 8.57]
Missing	4 (0.9%)
as.factor(Qscout.CO.COW)
0	366 (86.3%)
1	56 (13.2%)
Missing	2 (0.5%)
log.TOTAL.NEU.CO.COW
Mean (SD)	3.68 (1.27)
Median [Min, Max]	3.40 [1.39, 8.95]
Missing	2 (0.5%)
log.TOTAL.LYM.CO.COW
Mean (SD)	2.64 (1.18)
Median [Min, Max]	2.48 [0, 7.12]
Missing	2 (0.5%)
log.TOTAL.MAC.CO.COW
Mean (SD)	2.67 (1.34)
Median [Min, Max]	2.40 [0, 7.61]
Missing	2 (0.5%)
PERCENT.NEU.CO.COW
Mean (SD)	56.2 (11.1)
Median [Min, Max]	56.5 [18.8, 86.7]
Missing	2 (0.5%)
PERCENT.LYM.CO.COW
Mean (SD)	20.8 (7.24)
Median [Min, Max]	20.0 [0, 44.8]
Missing	2 (0.5%)
PERCENT.MAC.CO.COW
Mean (SD)	23.0 (11.7)
Median [Min, Max]	21.2 [2.00, 71.9]
Missing	2 (0.5%)

Best cut-off point Dry-off

Cow IMI

# Best cut-off using ROC analysis

# Set a seed for reproducibility

set.seed(123)

# Dry-off

# ------ Cow level -------

# Reference IMI

# SCC IMI COW

SCC.IMI.CO <- cutpointr(Doff.by.cow.3, log.SCC.CO.COW, IMI.cow, 
                     method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(SCC.IMI.CO)

## Method: maximize_metric 
## Predictor: log.SCC.CO.COW 
## Outcome: IMI.cow 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC   n n_pos n_neg
##  0.5847 480   382    98
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity   ppv    npv  tp
##            4.7095        1.1886 0.5354      0.4948      0.6939 0.863 0.2605 189
##   fn fp tn
##  193 30 68
## 
## Predictor summary: 
##     Data     Min.       5%  1st Qu.   Median     Mean  3rd Qu.      95%    Max.
##  Overall 1.791759 2.944439 3.912023 4.564348 4.681304 5.399246 6.877693 8.05484
##        0 1.945910 2.988038 3.719596 4.369127 4.408228 4.998861 6.002250 8.05484
##        1 1.791759 2.947004 3.956006 4.691348 4.751360 5.499209 7.009273 7.99699
##         SD NAs
##  1.1466263   0
##  0.9963544   0
##  1.1730676   0
## 
## Bootstrap summary: 
##           Variable  Min.   5% 1st Qu. Median Mean 3rd Qu.  95% Max.   SD NAs
##   optimal_cutpoint  3.53 4.28    4.71   4.71 4.77    4.74 5.55 6.28 0.34   0
##              AUC_b  0.47 0.54    0.56   0.59 0.59    0.61 0.64 0.68 0.03   0
##            AUC_oob  0.43 0.52    0.56   0.58 0.58    0.61 0.65 0.70 0.04   0
##    sum_sens_spec_b  1.02 1.12    1.16   1.20 1.20    1.23 1.28 1.39 0.05   0
##  sum_sens_spec_oob  0.88 1.00    1.08   1.14 1.14    1.19 1.26 1.37 0.08   0
##              acc_b  0.24 0.39    0.52   0.54 0.53    0.56 0.63 0.78 0.07   0
##            acc_oob  0.25 0.36    0.49   0.52 0.51    0.55 0.60 0.73 0.07   0
##      sensitivity_b  0.09 0.25    0.46   0.49 0.48    0.52 0.66 0.89 0.11   0
##    sensitivity_oob  0.06 0.23    0.44   0.48 0.47    0.51 0.65 0.85 0.11   0
##      specificity_b  0.22 0.54    0.68   0.72 0.72    0.76 0.90 1.00 0.10   0
##    specificity_oob  0.14 0.41    0.62   0.68 0.67    0.74 0.87 1.00 0.13   0
##     cohens_kappa_b  0.01 0.07    0.10   0.12 0.12    0.15 0.18 0.26 0.03   0
##   cohens_kappa_oob -0.10 0.00    0.05   0.08 0.08    0.12 0.16 0.25 0.05   0
##              ppv_b  0.78 0.83    0.85   0.87 0.87    0.89 0.92 1.00 0.03   0
##            ppv_oob  0.71 0.79    0.82   0.85 0.85    0.88 0.91 1.00 0.04   0
##              npv_b  0.17 0.22    0.25   0.26 0.26    0.28 0.31 0.40 0.03   0
##            npv_oob  0.12 0.19    0.22   0.24 0.24    0.27 0.30 0.38 0.04   0

# ROC plot
plot(SCC.IMI.CO)

# Youden index plot
plot_metric(SCC.IMI.CO)

# NEU IMI COW
NEU.IMI.CO <- cutpointr(Doff.by.cow.3, log.TOTAL.NEU.CO.COW, IMI.cow, 
                        method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(NEU.IMI.CO)

## Method: maximize_metric 
## Predictor: log.TOTAL.NEU.CO.COW 
## Outcome: IMI.cow 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC   n n_pos n_neg
##  0.6038 468   371    97
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv
##            4.2767          1.19 0.4786      0.3962      0.7938 0.8802 0.2558
##   tp  fn fp tn
##  147 224 20 77
## 
## Predictor summary: 
##     Data     Min.       5%  1st Qu.   Median     Mean  3rd Qu.      95%
##  Overall 1.386294 2.302585 3.124381 3.839394 4.025560 4.736198 6.424920
##        0 1.791759 2.448442 2.890372 3.433987 3.664973 4.204693 5.304821
##        1 1.386294 2.302585 3.218876 3.931826 4.119838 4.770685 6.810725
##      Max.       SD NAs
##  8.674026 1.286657   0
##  8.176673 1.018763   0
##  8.674026 1.333119   0
## 
## Bootstrap summary: 
##           Variable  Min.   5% 1st Qu. Median Mean 3rd Qu.  95% Max.   SD NAs
##   optimal_cutpoint  3.00 3.09    3.40   4.06 3.84    4.28 4.60 5.95 0.51   0
##              AUC_b  0.50 0.55    0.58   0.60 0.60    0.62 0.65 0.69 0.03   0
##            AUC_oob  0.47 0.53    0.58   0.61 0.60    0.63 0.67 0.72 0.04   0
##    sum_sens_spec_b  1.05 1.14    1.19   1.22 1.22    1.25 1.29 1.37 0.05   0
##  sum_sens_spec_oob  0.87 1.02    1.09   1.13 1.13    1.18 1.24 1.37 0.07   0
##              acc_b  0.27 0.44    0.49   0.56 0.58    0.67 0.73 0.78 0.10   0
##            acc_oob  0.24 0.41    0.46   0.53 0.55    0.64 0.71 0.78 0.10   0
##      sensitivity_b  0.09 0.34    0.41   0.50 0.56    0.71 0.81 0.87 0.17   0
##    sensitivity_oob  0.06 0.31    0.39   0.49 0.54    0.70 0.81 0.86 0.17   0
##      specificity_b  0.28 0.39    0.51   0.70 0.66    0.81 0.87 1.00 0.17   0
##    specificity_oob  0.13 0.31    0.42   0.59 0.59    0.76 0.85 1.00 0.19   0
##     cohens_kappa_b  0.03 0.08    0.12   0.15 0.15    0.18 0.23 0.35 0.05   0
##   cohens_kappa_oob -0.11 0.01    0.06   0.09 0.09    0.12 0.18 0.31 0.05   0
##              ppv_b  0.78 0.82    0.85   0.87 0.87    0.89 0.92 1.00 0.03   0
##            ppv_oob  0.72 0.78    0.82   0.84 0.84    0.87 0.91 1.00 0.04   0
##              npv_b  0.19 0.23    0.26   0.29 0.29    0.32 0.37 0.47 0.04   0
##            npv_oob  0.11 0.19    0.23   0.25 0.26    0.28 0.34 0.49 0.04   0

# ROC plot
plot(NEU.IMI.CO)

# Youden index plot
plot_metric(NEU.IMI.CO)

# %NEU IMI COW
PER.NEU.IMI.CO <- cutpointr(Doff.by.cow.3, PERCENT.NEU.CO.COW, IMI.cow, 
                        method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(PER.NEU.IMI.CO)

## Method: maximize_metric 
## Predictor: PERCENT.NEU.CO.COW 
## Outcome: IMI.cow 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC   n n_pos n_neg
##  0.5488 468   371    97
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv   npv  tp
##             51.63        1.1329 0.5299       0.504      0.6289 0.8386 0.249 187
##   fn fp tn
##  184 36 61
## 
## Predictor summary: 
##     Data  Min.      5% 1st Qu. Median     Mean 3rd Qu.    95%   Max.       SD
##  Overall 15.69 29.2245 41.6575  51.14 50.68026   59.58 70.399 122.00 12.80270
##        0 19.90 29.2000 41.4800  49.33 49.41639   56.64 68.902 122.00 13.75413
##        1 15.69 29.2650 41.6850  51.69 51.01070   60.82 71.055  83.52 12.54066
##  NAs
##    0
##    0
##    0
## 
## Bootstrap summary: 
##           Variable  Min.    5% 1st Qu. Median  Mean 3rd Qu.   95%  Max.   SD
##   optimal_cutpoint 31.23 45.90   51.63  51.79 53.55   57.63 62.38 65.30 5.49
##              AUC_b  0.44  0.50    0.53   0.55  0.55    0.57  0.60  0.65 0.03
##            AUC_oob  0.42  0.48    0.52   0.55  0.55    0.58  0.62  0.69 0.04
##    sum_sens_spec_b  1.02  1.08    1.12   1.15  1.15    1.18  1.23  1.33 0.05
##  sum_sens_spec_oob  0.80  0.95    1.02   1.07  1.07    1.12  1.18  1.33 0.07
##              acc_b  0.28  0.35    0.43   0.53  0.50    0.55  0.62  0.80 0.09
##            acc_oob  0.24  0.33    0.40   0.50  0.48    0.53  0.58  0.79 0.09
##      sensitivity_b  0.12  0.21    0.33   0.49  0.45    0.53  0.65  0.95 0.15
##    sensitivity_oob  0.10  0.19    0.30   0.47  0.44    0.52  0.63  0.96 0.15
##      specificity_b  0.13  0.49    0.63   0.68  0.70    0.81  0.91  0.98 0.14
##    specificity_oob  0.03  0.34    0.55   0.62  0.63    0.74  0.89  0.97 0.17
##     cohens_kappa_b  0.01  0.04    0.07   0.09  0.09    0.12  0.15  0.22 0.03
##   cohens_kappa_oob -0.18 -0.04    0.01   0.04  0.04    0.07  0.12  0.22 0.05
##              ppv_b  0.76  0.81    0.84   0.86  0.86    0.88  0.91  0.97 0.03
##            ppv_oob  0.70  0.76    0.80   0.82  0.82    0.85  0.89  0.97 0.04
##              npv_b  0.18  0.21    0.23   0.25  0.25    0.27  0.31  0.46 0.03
##            npv_oob  0.03  0.17    0.20   0.23  0.23    0.25  0.28  0.35 0.04
##  NAs
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0

# ROC plot
plot(PER.NEU.IMI.CO)

# Youden index plot
plot_metric(PER.NEU.IMI.CO)

# LYM IMI COW
LYM.IMI.CO  <- cutpointr(Doff.by.cow.3, log.TOTAL.LYM.CO.COW, IMI.cow, 
                        method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000)  %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(LYM.IMI.CO)

## Method: maximize_metric 
## Predictor: log.TOTAL.LYM.CO.COW 
## Outcome: IMI.cow 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC   n n_pos n_neg
##  0.6116 468   371    97
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv
##            2.7726        1.2209 0.5876      0.5714      0.6495 0.8618 0.2838
##   tp  fn fp tn
##  212 159 34 63
## 
## Predictor summary: 
##     Data      Min.       5%  1st Qu.   Median     Mean  3rd Qu.      95%
##  Overall 0.0000000 1.098612 2.079442 2.802901 2.918222 3.644080 4.904848
##        0 0.6931472 1.098612 1.945910 2.484907 2.593755 3.218876 4.256046
##        1 0.0000000 1.098612 2.197225 2.944439 3.003056 3.749435 5.100869
##      Max.        SD NAs
##  7.082549 1.1648283   0
##  6.786717 0.9968735   0
##  7.082549 1.1915860   0
## 
## Bootstrap summary: 
##           Variable  Min.   5% 1st Qu. Median Mean 3rd Qu.  95% Max.   SD NAs
##   optimal_cutpoint  2.08 2.64    2.64   2.71 2.76    2.77 3.53 4.50 0.27   0
##              AUC_b  0.52 0.56    0.59   0.61 0.61    0.63 0.66 0.71 0.03   0
##            AUC_oob  0.45 0.54    0.58   0.61 0.61    0.64 0.68 0.76 0.04   0
##    sum_sens_spec_b  1.07 1.15    1.20   1.23 1.24    1.27 1.32 1.42 0.05   0
##  sum_sens_spec_oob  0.92 1.05    1.13   1.19 1.18    1.24 1.31 1.44 0.08   0
##              acc_b  0.29 0.44    0.59   0.61 0.60    0.63 0.66 0.78 0.06   0
##            acc_oob  0.28 0.41    0.57   0.59 0.58    0.62 0.65 0.75 0.06   0
##      sensitivity_b  0.12 0.34    0.57   0.60 0.59    0.63 0.68 0.86 0.09   0
##    sensitivity_oob  0.10 0.30    0.56   0.59 0.58    0.63 0.68 0.85 0.10   0
##      specificity_b  0.34 0.53    0.60   0.64 0.65    0.68 0.86 1.00 0.09   0
##    specificity_oob  0.16 0.45    0.55   0.61 0.61    0.67 0.81 1.00 0.11   0
##     cohens_kappa_b  0.04 0.09    0.14   0.17 0.16    0.19 0.23 0.36 0.04   0
##   cohens_kappa_oob -0.08 0.03    0.09   0.13 0.13    0.17 0.22 0.30 0.06   0
##              ppv_b  0.79 0.83    0.85   0.87 0.87    0.88 0.91 1.00 0.03   0
##            ppv_oob  0.73 0.80    0.83   0.85 0.85    0.87 0.90 1.00 0.03   0
##              npv_b  0.17 0.24    0.27   0.29 0.29    0.32 0.35 0.50 0.03   0
##            npv_oob  0.15 0.21    0.24   0.27 0.27    0.30 0.34 0.41 0.04   0

# ROC plot
plot(LYM.IMI.CO)

# Youden index plot
plot_metric(LYM.IMI.CO)

# %LYM IMI COW
PER.LYM.IMI.CO <- cutpointr(Doff.by.cow.3, PERCENT.LYM.CO.COW, IMI.cow, 
                            method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(PER.LYM.IMI.CO)

## Method: maximize_metric 
## Predictor: PERCENT.LYM.CO.COW 
## Outcome: IMI.cow 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC   n n_pos n_neg
##  0.5002 468   371    97
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv
##              15.2        1.0596 0.5684      0.5957      0.4639 0.8095 0.2308
##   tp  fn fp tn
##  221 150 52 45
## 
## Predictor summary: 
##     Data Min.     5% 1st Qu. Median     Mean 3rd Qu.    95% Max.       SD NAs
##  Overall 3.54 8.0935 12.7075  16.88 17.64716  21.125 30.360 69.0 7.389449   0
##        0 4.94 8.6400 12.7500  16.67 17.91680  20.600 31.104 69.0 8.508381   0
##        1 3.54 8.0950 12.7050  16.89 17.57666  21.140 29.915 54.5 7.078988   0
## 
## Bootstrap summary: 
##           Variable  Min.    5% 1st Qu. Median  Mean 3rd Qu.   95% Max.   SD NAs
##   optimal_cutpoint  5.56  9.07   14.61  15.20   Inf   15.79 24.00  Inf  NaN   0
##              AUC_b  0.38  0.45    0.48   0.50  0.50    0.52  0.56 0.60 0.03   0
##            AUC_oob  0.36  0.43    0.47   0.50  0.50    0.53  0.57 0.67 0.04   0
##    sum_sens_spec_b  0.94  1.02    1.05   1.08  1.09    1.12  1.17 1.25 0.05   0
##  sum_sens_spec_oob  0.72  0.88    0.94   0.98  0.98    1.02  1.09 1.22 0.07   0
##              acc_b  0.18  0.32    0.54   0.58  0.57    0.62  0.78 0.82 0.13   0
##            acc_oob  0.15  0.27    0.50   0.55  0.53    0.59  0.76 0.83 0.14   0
##      sensitivity_b  0.00  0.17    0.55   0.60  0.58    0.66  0.95 1.00 0.23   0
##    sensitivity_oob  0.00  0.14    0.51   0.58  0.56    0.64  0.94 1.00 0.23   0
##      specificity_b  0.01  0.10    0.42   0.50  0.50    0.58  0.89 1.00 0.23   0
##    specificity_oob  0.00  0.03    0.31   0.42  0.42    0.52  0.83 1.00 0.23   0
##     cohens_kappa_b -0.06  0.01    0.04   0.06  0.06    0.09  0.13 0.19 0.03   0
##   cohens_kappa_oob -0.23 -0.09   -0.04  -0.01 -0.01    0.02  0.07 0.14 0.05   0
##              ppv_b  0.74  0.78    0.80   0.82  0.83    0.84  0.88 1.00 0.03   4
##            ppv_oob  0.00  0.71    0.76   0.79  0.78    0.81  0.84 1.00 0.06   5
##              npv_b  0.15  0.20    0.23   0.25  0.26    0.27  0.35 1.00 0.06   0
##            npv_oob  0.00  0.11    0.17   0.20  0.19    0.22  0.26 0.45 0.05   8

# ROC plot
plot(PER.LYM.IMI.CO)

## Warning: Removed 4 rows containing non-finite outside the scale range
## (`stat_density()`).

# Youden index plot
plot_metric(PER.LYM.IMI.CO)

# MAC IMI COW
MAC.IMI.CO <- cutpointr(Doff.by.cow.3, log.TOTAL.MAC.CO.COW, IMI.cow, 
                        method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(MAC.IMI.CO)

## Method: maximize_metric 
## Predictor: log.TOTAL.MAC.CO.COW 
## Outcome: IMI.cow 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC   n n_pos n_neg
##  0.5733 468   371    97
## 
##  optimal_cutpoint sum_sens_spec   acc sensitivity specificity    ppv    npv  tp
##            4.0604        1.1388 0.438       0.345      0.7938 0.8649 0.2406 128
##   fn fp tn
##  243 20 77
## 
## Predictor summary: 
##     Data Min.       5%  1st Qu.   Median     Mean  3rd Qu.      95%     Max.
##  Overall    0 1.609438 2.564949 3.401197 3.484217 4.406719 5.743569 7.289611
##        0    0 1.328758 2.197225 3.135494 3.211078 3.988984 5.257283 6.419995
##        1    0 1.609438 2.564949 3.433987 3.555631 4.521789 5.791487 7.289611
##        SD NAs
##  1.335529   0
##  1.226225   0
##  1.355151   0
## 
## Bootstrap summary: 
##           Variable  Min.    5% 1st Qu. Median Mean 3rd Qu.  95% Max.   SD NAs
##   optimal_cutpoint  1.79  2.64    3.22   3.74 3.75    4.14 4.61 5.32 0.56   0
##              AUC_b  0.47  0.52    0.55   0.57 0.57    0.59 0.62 0.68 0.03   0
##            AUC_oob  0.45  0.51    0.54   0.57 0.57    0.60 0.64 0.72 0.04   0
##    sum_sens_spec_b  1.04  1.09    1.14   1.17 1.17    1.20 1.25 1.33 0.05   0
##  sum_sens_spec_oob  0.79  0.97    1.04   1.08 1.08    1.13 1.19 1.32 0.07   0
##              acc_b  0.28  0.37    0.44   0.49 0.51    0.58 0.66 0.78 0.09   0
##            acc_oob  0.21  0.35    0.41   0.46 0.47    0.54 0.63 0.74 0.09   0
##      sensitivity_b  0.11  0.24    0.35   0.42 0.45    0.57 0.74 0.93 0.15   0
##    sensitivity_oob  0.06  0.22    0.32   0.39 0.43    0.55 0.71 0.92 0.15   0
##      specificity_b  0.16  0.41    0.62   0.76 0.72    0.83 0.90 0.99 0.15   0
##    specificity_oob  0.05  0.29    0.54   0.69 0.65    0.78 0.88 1.00 0.17   0
##     cohens_kappa_b  0.02  0.05    0.08   0.10 0.10    0.13 0.17 0.23 0.03   0
##   cohens_kappa_oob -0.16 -0.02    0.02   0.05 0.05    0.07 0.12 0.20 0.04   0
##              ppv_b  0.76  0.82    0.85   0.87 0.87    0.89 0.92 0.99 0.03   0
##            ppv_oob  0.68  0.77    0.80   0.83 0.83    0.86 0.90 1.00 0.04   0
##              npv_b  0.16  0.21    0.24   0.26 0.26    0.28 0.31 0.45 0.03   0
##            npv_oob  0.09  0.17    0.21   0.23 0.23    0.25 0.28 0.38 0.03   0

# ROC plot
plot(MAC.IMI.CO)

# Youden index plot
plot_metric(MAC.IMI.CO)

# %MAC IMI COW
PER.MAC.IMI.CO <- cutpointr(Doff.by.cow.3, PERCENT.MAC.CO.COW, IMI.cow, 
                            method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 0

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(PER.MAC.IMI.CO)

## Method: maximize_metric 
## Predictor: PERCENT.MAC.CO.COW 
## Outcome: IMI.cow 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC   n n_pos n_neg
##  0.5488 468    97   371
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv tp
##             43.55        1.0995 0.6966       0.299      0.8005 0.2816 0.8137 29
##  fn fp  tn
##  68 74 297
## 
## Predictor summary: 
##     Data Min.      5% 1st Qu. Median     Mean 3rd Qu.     95%  Max.       SD
##  Overall 3.41 12.5175 20.8225  29.93 31.98408   41.53 58.0975 74.51 14.27458
##        0 4.76 12.6560 21.5000  31.11 34.16928   44.59 60.4840 72.50 15.35945
##        1 3.41 12.5250 20.2800  29.55 31.41275   41.24 55.3750 74.51 13.94252
##  NAs
##    0
##    0
##    0
## 
## Bootstrap summary: 
##           Variable  Min.    5% 1st Qu. Median  Mean 3rd Qu.   95%  Max.    SD
##   optimal_cutpoint 10.00 19.12   27.80  36.84 36.23   43.94 52.78 65.80 10.22
##              AUC_b  0.43  0.49    0.53   0.55  0.55    0.57  0.60  0.66  0.03
##            AUC_oob  0.40  0.48    0.52   0.55  0.55    0.58  0.62  0.70  0.04
##    sum_sens_spec_b  0.99  1.07    1.10   1.13  1.13    1.16  1.21  1.30  0.04
##  sum_sens_spec_oob  0.76  0.92    0.98   1.02  1.02    1.07  1.12  1.23  0.06
##              acc_b  0.19  0.37    0.51   0.63  0.61    0.72  0.78  0.83  0.13
##            acc_oob  0.23  0.33    0.47   0.58  0.57    0.69  0.75  0.82  0.13
##      sensitivity_b  0.05  0.18    0.31   0.47  0.50    0.66  0.88  1.00  0.22
##    sensitivity_oob  0.00  0.10    0.23   0.38  0.41    0.56  0.81  0.97  0.22
##      specificity_b  0.03  0.24    0.48   0.67  0.64    0.82  0.93  0.99  0.22
##    specificity_oob  0.02  0.21    0.45   0.63  0.62    0.81  0.92  0.99  0.22
##     cohens_kappa_b  0.00  0.04    0.08   0.10  0.11    0.14  0.19  0.27  0.04
##   cohens_kappa_oob -0.16 -0.06   -0.01   0.02  0.02    0.05  0.11  0.21  0.05
##              ppv_b  0.16  0.21    0.24   0.28  0.29    0.32  0.41  0.73  0.06
##            ppv_oob  0.00  0.15    0.19   0.22  0.23    0.25  0.32  0.56  0.06
##              npv_b  0.76  0.79    0.82   0.83  0.84    0.85  0.89  1.00  0.03
##            npv_oob  0.62  0.74    0.78   0.80  0.80    0.82  0.85  0.94  0.03
##  NAs
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0

# ROC plot
plot(PER.MAC.IMI.CO)

# Youden index plot
plot_metric(PER.MAC.IMI.CO)

# Summary of results when using IMI as reference

print(c("Optimal SCC:", exp(SCC.IMI.CO$optimal_cutpoint), SCC.IMI.CO$sensitivity, SCC.IMI.CO$specificity, SCC.IMI.CO$ppv, SCC.IMI.CO$npv, SCC.IMI.CO$AUC))

##                                                 sensitivity         specificity 
##      "Optimal SCC:"               "111" "0.494764397905759" "0.693877551020408" 
##                                                             
## "0.863013698630137" "0.260536398467433" "0.584731274708836"

print(c("Optimal n NEU:", exp(NEU.IMI.CO$optimal_cutpoint), NEU.IMI.CO$sensitivity, NEU.IMI.CO$specificity, NEU.IMI.CO$ppv, NEU.IMI.CO$npv, NEU.IMI.CO$AUC))

##                                                 sensitivity         specificity 
##    "Optimal n NEU:"                "72"  "0.39622641509434" "0.793814432989691" 
##                                                             
## "0.880239520958084" "0.255813953488372" "0.603815266624059"

print(c("Optimal %NEU:", PER.NEU.IMI.CO$optimal_cutpoint, PER.NEU.IMI.CO$sensitivity, PER.NEU.IMI.CO$specificity, PER.NEU.IMI.CO$ppv, PER.NEU.IMI.CO$npv, PER.NEU.IMI.CO$AUC))

##                                                 sensitivity         specificity 
##     "Optimal %NEU:"             "51.63" "0.504043126684636" "0.628865979381443" 
##                                                             
## "0.838565022421525" "0.248979591836735" "0.548823186150554"

print(c("Optimal n LYM:", exp(LYM.IMI.CO$optimal_cutpoint), LYM.IMI.CO$sensitivity, LYM.IMI.CO$specificity, LYM.IMI.CO$ppv, LYM.IMI.CO$npv, LYM.IMI.CO$AUC))

##                                                 sensitivity         specificity 
##    "Optimal n LYM:"                "16" "0.571428571428571" "0.649484536082474" 
##                                                             
## "0.861788617886179" "0.283783783783784" "0.611568066246144"

print(c("Optimal %LYM:", PER.LYM.IMI.CO$optimal_cutpoint, PER.LYM.IMI.CO$sensitivity, PER.LYM.IMI.CO$specificity, PER.LYM.IMI.CO$ppv, PER.LYM.IMI.CO$npv, PER.LYM.IMI.CO$AUC))

##                                                 sensitivity         specificity 
##     "Optimal %LYM:"              "15.2" "0.595687331536388" "0.463917525773196" 
##                                                             
##  "0.80952380952381" "0.230769230769231"  "0.50015283296746"

print(c("Optimal n MAC:", exp(MAC.IMI.CO$optimal_cutpoint), MAC.IMI.CO$sensitivity, MAC.IMI.CO$specificity, MAC.IMI.CO$ppv, MAC.IMI.CO$npv, MAC.IMI.CO$AUC))

##                                                 sensitivity         specificity 
##    "Optimal n MAC:"                "58" "0.345013477088949" "0.793814432989691" 
##                                                             
## "0.864864864864865"          "0.240625" "0.573304248756495"

print(c("Optimal %MAC:", PER.MAC.IMI.CO$optimal_cutpoint, PER.MAC.IMI.CO$sensitivity, PER.MAC.IMI.CO$specificity, PER.MAC.IMI.CO$ppv, PER.MAC.IMI.CO$npv, PER.MAC.IMI.CO$AUC))

##                                                 sensitivity         specificity 
##     "Optimal %MAC:"             "43.55" "0.298969072164948" "0.800539083557951" 
##                                                             
## "0.281553398058252" "0.813698630136986" "0.548795398338289"

# CI using bp

# SCC

SCC.IMI.CO_optimal_cutpoint <- boot_ci(SCC.IMI.CO, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
SCC.IMI.CO_lower_ci <- subset(SCC.IMI.CO_optimal_cutpoint,SCC.IMI.CO_optimal_cutpoint$quantile==0.025)$values
exp(SCC.IMI.CO_lower_ci)

## [1] 64.92326

SCC.IMI.CO_upper_ci <- subset(SCC.IMI.CO_optimal_cutpoint,SCC.IMI.CO_optimal_cutpoint$quantile==0.975)$values
exp(SCC.IMI.CO_upper_ci)

## [1] 279

boot_ci(SCC.IMI.CO, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.220
## 2    0.975  0.701

boot_ci(SCC.IMI.CO, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.465
## 2    0.975  0.931

boot_ci(SCC.IMI.CO, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.822
## 2    0.975  0.931

boot_ci(SCC.IMI.CO, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.208
## 2    0.975  0.324

boot_ci(SCC.IMI.CO, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.526
## 2    0.975  0.645

# Number neu

NEU.IMI.CO_optimal_cutpoint <- boot_ci(NEU.IMI.CO, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)

NEU.IMI.CO_lower_ci <- subset(NEU.IMI.CO_optimal_cutpoint,NEU.IMI.CO_optimal_cutpoint$quantile==0.025)$values
exp(NEU.IMI.CO_lower_ci)

## [1] 22

NEU.IMI.CO_upper_ci <- subset(NEU.IMI.CO_optimal_cutpoint,NEU.IMI.CO_optimal_cutpoint$quantile==0.975)$values
exp(NEU.IMI.CO_upper_ci)

## [1] 103

boot_ci(NEU.IMI.CO, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.310
## 2    0.975  0.826

boot_ci(NEU.IMI.CO, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.370
## 2    0.975  0.892

boot_ci(NEU.IMI.CO, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.813
## 2    0.975  0.933

boot_ci(NEU.IMI.CO, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.220
## 2    0.975  0.386

boot_ci(NEU.IMI.CO, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.544
## 2    0.975  0.661

# Percent neu

PER.NEU.IMI.CO_optimal_cutpoint <- boot_ci(PER.NEU.IMI.CO, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
PER.NEU.IMI.CO_optimal_cutpoint

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025   36.8
## 2    0.975   64.0

PER.NEU.IMI.CO_lower_ci <- subset(PER.NEU.IMI.CO_optimal_cutpoint,PER.NEU.IMI.CO_optimal_cutpoint$quantile==0.025)$values
PER.NEU.IMI.CO_upper_ci <- subset(PER.NEU.IMI.CO_optimal_cutpoint,PER.NEU.IMI.CO_optimal_cutpoint$quantile==0.975)$values


boot_ci(PER.NEU.IMI.CO, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.192
## 2    0.975  0.876

boot_ci(PER.NEU.IMI.CO, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.257
## 2    0.975  0.931

boot_ci(PER.NEU.IMI.CO, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.802
## 2    0.975  0.928

boot_ci(PER.NEU.IMI.CO, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.201
## 2    0.975  0.329

boot_ci(PER.NEU.IMI.CO, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.488
## 2    0.975  0.613

# Number lym

LYM.IMI.CO_optimal_cutpoint <- boot_ci(LYM.IMI.CO, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)

LYM.IMI.CO_lower_ci <- subset(LYM.IMI.CO_optimal_cutpoint,LYM.IMI.CO_optimal_cutpoint$quantile==0.025)$values
exp(LYM.IMI.CO_lower_ci)

## [1] 11

LYM.IMI.CO_upper_ci <- subset(LYM.IMI.CO_optimal_cutpoint,LYM.IMI.CO_optimal_cutpoint$quantile==0.975)$values
exp(LYM.IMI.CO_upper_ci)

## [1] 41

boot_ci(LYM.IMI.CO, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.268
## 2    0.975  0.710

boot_ci(LYM.IMI.CO, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.483
## 2    0.975  0.913

boot_ci(LYM.IMI.CO, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.821
## 2    0.975  0.922

boot_ci(LYM.IMI.CO, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.228
## 2    0.975  0.362

boot_ci(LYM.IMI.CO, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.553
## 2    0.975  0.674

# Percent lym

PER.LYM.IMI.CO_optimal_cutpoint <- boot_ci(PER.LYM.IMI.CO, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
PER.LYM.IMI.CO_optimal_cutpoint

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025   7.79
## 2    0.975  31.4

PER.LYM.IMI.CO_lower_ci <- subset(PER.LYM.IMI.CO_optimal_cutpoint,PER.LYM.IMI.CO_optimal_cutpoint$quantile==0.025)$values
PER.LYM.IMI.CO_upper_ci <- subset(PER.LYM.IMI.CO_optimal_cutpoint,PER.LYM.IMI.CO_optimal_cutpoint$quantile==0.975)$values

boot_ci(PER.LYM.IMI.CO, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025 0.0413
## 2    0.975 0.968

boot_ci(PER.LYM.IMI.CO, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025 0.0638
## 2    0.975 0.979

boot_ci(PER.LYM.IMI.CO, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.774
## 2    0.975  0.905

boot_ci(PER.LYM.IMI.CO, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.193
## 2    0.975  0.4

boot_ci(PER.LYM.IMI.CO, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.433
## 2    0.975  0.567

# Number mac

MAC.IMI.CO_optimal_cutpoint <- boot_ci(MAC.IMI.CO, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
MAC.IMI.CO_lower_ci <- subset(MAC.IMI.CO_optimal_cutpoint,MAC.IMI.CO_optimal_cutpoint$quantile==0.025)$values
exp(MAC.IMI.CO_lower_ci)

## [1] 10

MAC.IMI.CO_upper_ci <- subset(MAC.IMI.CO_optimal_cutpoint,MAC.IMI.CO_optimal_cutpoint$quantile==0.975)$values
exp(MAC.IMI.CO_upper_ci)

## [1] 124

boot_ci(MAC.IMI.CO, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.214
## 2    0.975  0.816

boot_ci(MAC.IMI.CO, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.330
## 2    0.975  0.925

boot_ci(MAC.IMI.CO, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.808
## 2    0.975  0.929

boot_ci(MAC.IMI.CO, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.205
## 2    0.975  0.329

boot_ci(MAC.IMI.CO, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.507
## 2    0.975  0.633

# Percent mac

PER.MAC.IMI.CO_optimal_cutpoint <- boot_ci(PER.MAC.IMI.CO, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
PER.MAC.IMI.CO_optimal_cutpoint

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025   19.1
## 2    0.975   55

PER.MAC.IMI.CO_lower_ci <- subset(PER.MAC.IMI.CO_optimal_cutpoint,PER.MAC.IMI.CO_optimal_cutpoint$quantile==0.025)$values
PER.MAC.IMI.CO_upper_ci <- subset(PER.MAC.IMI.CO_optimal_cutpoint,PER.MAC.IMI.CO_optimal_cutpoint$quantile==0.975)$values

boot_ci(PER.MAC.IMI.CO, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.139
## 2    0.975  0.898

boot_ci(PER.MAC.IMI.CO, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.216
## 2    0.975  0.947

boot_ci(PER.MAC.IMI.CO, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.205
## 2    0.975  0.441

boot_ci(PER.MAC.IMI.CO, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.786
## 2    0.975  0.900

boot_ci(PER.MAC.IMI.CO, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.482
## 2    0.975  0.615

# Prevalence

# True prevalence

table(Doff.by.cow.3$IMI.cow)

## 
##   0   1 
##  98 384

prop.table(table(Doff.by.cow.3$IMI.cow))

## 
##         0         1 
## 0.2033195 0.7966805

#Apparent prevalence

# ---- SCC ----

# optimal cut-off

optimal_scc_doff_IMI_cow <- SCC.IMI.CO$optimal_cutpoint

table(Doff.by.cow.3$SCC.CO.COW > exp(optimal_scc_doff_IMI_cow))

## 
## FALSE  TRUE 
##   264   222

prop.table(table(Doff.by.cow.3$SCC.CO.COW > exp(optimal_scc_doff_IMI_cow)))

## 
##     FALSE      TRUE 
## 0.5432099 0.4567901

# confidence intervals

table(Doff.by.cow.3$SCC.CO.COW > exp(SCC.IMI.CO_lower_ci))

## 
## FALSE  TRUE 
##   168   318

prop.table(table(Doff.by.cow.3$SCC.CO.COW > exp(SCC.IMI.CO_lower_ci)))

## 
##    FALSE     TRUE 
## 0.345679 0.654321

table(Doff.by.cow.3$SCC.CO.COW > exp(SCC.IMI.CO_upper_ci))

## 
## FALSE  TRUE 
##   392    94

prop.table(table(Doff.by.cow.3$SCC.CO.COW >exp(SCC.IMI.CO_upper_ci)))

## 
##     FALSE      TRUE 
## 0.8065844 0.1934156

# ---- Neutrophils ----

# optimal cut-off

optimal_neu_doff_IMI_cow <- NEU.IMI.CO$optimal_cutpoint

table(Doff.by.cow.3$TOTAL.NEU.CO.COW > exp(optimal_neu_doff_IMI_cow))

## 
## FALSE  TRUE 
##   304   169

prop.table(table(Doff.by.cow.3$TOTAL.NEU.CO.COW > exp(optimal_neu_doff_IMI_cow)))

## 
##     FALSE      TRUE 
## 0.6427061 0.3572939

# confidence intervals

table(Doff.by.cow.3$TOTAL.NEU.CO.COW > exp(NEU.IMI.CO_lower_ci))

## 
## FALSE  TRUE 
##   117   356

prop.table(table(Doff.by.cow.3$TOTAL.NEU.CO.COW > exp(NEU.IMI.CO_lower_ci)))

## 
##     FALSE      TRUE 
## 0.2473573 0.7526427

table(Doff.by.cow.3$TOTAL.NEU.CO.COW > exp(NEU.IMI.CO_upper_ci))

## 
## FALSE  TRUE 
##   342   131

prop.table(table(Doff.by.cow.3$TOTAL.NEU.CO.COW >exp(NEU.IMI.CO_upper_ci)))

## 
##     FALSE      TRUE 
## 0.7230444 0.2769556

# ---- Percent neu ----

# Optimal cut-off

optimal_per_neu_doff_IMI_cow <-PER.NEU.IMI.CO$optimal_cutpoint

table(Doff.by.cow.3$PERCENT.NEU.CO.COW > optimal_per_neu_doff_IMI_cow)

## 
## FALSE  TRUE 
##   249   224

prop.table(table(Doff.by.cow.3$PERCENT.NEU.CO.COW > optimal_per_neu_doff_IMI_cow))

## 
##     FALSE      TRUE 
## 0.5264271 0.4735729

# confidence intervals

table(Doff.by.cow.3$PERCENT.NEU.CO.COW > PER.NEU.IMI.CO_lower_ci)

## 
## FALSE  TRUE 
##    60   413

prop.table(table(Doff.by.cow.3$PERCENT.NEU.CO.COW > PER.NEU.IMI.CO_lower_ci))

## 
##     FALSE      TRUE 
## 0.1268499 0.8731501

table(Doff.by.cow.3$PERCENT.NEU.CO.COW > PER.NEU.IMI.CO_upper_ci)

## 
## FALSE  TRUE 
##   398    75

prop.table(table(Doff.by.cow.3$PERCENT.NEU.CO.COW >PER.NEU.IMI.CO_upper_ci))

## 
##     FALSE      TRUE 
## 0.8414376 0.1585624

# ---- Lymphocytes ----

# optimal cut-off

optimal_lym_doff_IMI_cow <-LYM.IMI.CO$optimal_cutpoint

table(Doff.by.cow.3$TOTAL.LYM.CO.COW > exp(optimal_lym_doff_IMI_cow))

## 
## FALSE  TRUE 
##   223   250

prop.table(table(Doff.by.cow.3$TOTAL.LYM.CO.COW > exp(optimal_lym_doff_IMI_cow)))

## 
##     FALSE      TRUE 
## 0.4714588 0.5285412

# confidence intervals

table(Doff.by.cow.3$TOTAL.LYM.CO.COW > exp(LYM.IMI.CO_lower_ci))

## 
## FALSE  TRUE 
##   175   298

prop.table(table(Doff.by.cow.3$TOTAL.LYM.CO.COW > exp(LYM.IMI.CO_lower_ci)))

## 
##     FALSE      TRUE 
## 0.3699789 0.6300211

table(Doff.by.cow.3$TOTAL.LYM.CO.COW > exp(LYM.IMI.CO_upper_ci))

## 
## FALSE  TRUE 
##   365   108

prop.table(table(Doff.by.cow.3$TOTAL.LYM.CO.COW >exp(LYM.IMI.CO_upper_ci)))

## 
##     FALSE      TRUE 
## 0.7716702 0.2283298

# ---- Percent lym ----

# optimal cut-off

optimal_per_lym_doff_IMI_cow <-PER.LYM.IMI.CO$optimal_cutpoint

table(Doff.by.cow.3$PERCENT.LYM.CO.COW > optimal_per_lym_doff_IMI_cow)

## 
## FALSE  TRUE 
##   197   276

prop.table(table(Doff.by.cow.3$PERCENT.LYM.CO.COW > optimal_per_lym_doff_IMI_cow))

## 
##     FALSE      TRUE 
## 0.4164905 0.5835095

# confidence intervals

table(Doff.by.cow.3$PERCENT.LYM.CO.COW > PER.LYM.IMI.CO_lower_ci)

## 
## FALSE  TRUE 
##    20   453

prop.table(table(Doff.by.cow.3$PERCENT.LYM.CO.COW > PER.LYM.IMI.CO_lower_ci))

## 
##     FALSE      TRUE 
## 0.0422833 0.9577167

table(Doff.by.cow.3$PERCENT.LYM.CO.COW > PER.LYM.IMI.CO_upper_ci)

## 
## FALSE  TRUE 
##   454    19

prop.table(table(Doff.by.cow.3$PERCENT.LYM.CO.COW >PER.LYM.IMI.CO_upper_ci))

## 
##      FALSE       TRUE 
## 0.95983087 0.04016913

# ---- Macrophages ----

# optimal cut-off

optimal_mac_doff_IMI_cow <- MAC.IMI.CO$optimal_cutpoint

table(Doff.by.cow.3$TOTAL.MAC.CO.COW > exp(optimal_mac_doff_IMI_cow))

## 
## FALSE  TRUE 
##   323   150

prop.table(table(Doff.by.cow.3$TOTAL.MAC.CO.COW > exp(optimal_mac_doff_IMI_cow)))

## 
##     FALSE      TRUE 
## 0.6828753 0.3171247

# confidence intervals

table(Doff.by.cow.3$TOTAL.MAC.CO.COW > exp(MAC.IMI.CO_lower_ci))

## 
## FALSE  TRUE 
##   101   372

prop.table(table(Doff.by.cow.3$TOTAL.MAC.CO.COW > exp(MAC.IMI.CO_lower_ci)))

## 
##     FALSE      TRUE 
## 0.2135307 0.7864693

table(Doff.by.cow.3$TOTAL.MAC.CO.COW > exp(MAC.IMI.CO_upper_ci))

## 
## FALSE  TRUE 
##   387    86

prop.table(table(Doff.by.cow.3$TOTAL.MAC.CO.COW >exp(MAC.IMI.CO_upper_ci)))

## 
##     FALSE      TRUE 
## 0.8181818 0.1818182

# Percent MAC

# optimal cut-off

optimal_per_mac_doff_IMI_cow <- PER.MAC.IMI.CO$optimal_cutpoint

table(Doff.by.cow.3$PERCENT.MAC.CO.COW > optimal_per_mac_doff_IMI_cow)

## 
## FALSE  TRUE 
##   371   102

prop.table(table(Doff.by.cow.3$PERCENT.MAC.CO.COW > optimal_per_mac_doff_IMI_cow))

## 
##     FALSE      TRUE 
## 0.7843552 0.2156448

# confidence intervals

table(Doff.by.cow.3$PERCENT.MAC.CO.COW > PER.MAC.IMI.CO_lower_ci)

## 
## FALSE  TRUE 
##   100   373

prop.table(table(Doff.by.cow.3$PERCENT.MAC.CO.COW > PER.MAC.IMI.CO_lower_ci))

## 
##     FALSE      TRUE 
## 0.2114165 0.7885835

table(Doff.by.cow.3$PERCENT.MAC.CO.COW > PER.MAC.IMI.CO_upper_ci)

## 
## FALSE  TRUE 
##   443    30

prop.table(table(Doff.by.cow.3$PERCENT.MAC.CO.COW >PER.MAC.IMI.CO_upper_ci))

## 
##      FALSE       TRUE 
## 0.93657505 0.06342495

Cow high SCC

# Set a seed for reproducibility

set.seed(123)

# Delaval SCM (SCC>200.000) reference

# n NEU SCM COW
NEU.SCM <- cutpointr(Doff.by.cow.3, log.TOTAL.NEU.CO.COW, delaval.SCM.CO.COW, 
                            method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(NEU.SCM)

## Method: maximize_metric 
## Predictor: log.TOTAL.NEU.CO.COW 
## Outcome: delaval.SCM.CO.COW 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC   n n_pos n_neg
##  0.9652 472   142   330
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv
##            4.2767        1.8346 0.9068      0.9437      0.8909 0.7882 0.9735
##   tp fn fp  tn
##  134  8 36 294
## 
## Predictor summary: 
##     Data     Min.       5%  1st Qu.   Median     Mean  3rd Qu.      95%
##  Overall 1.386294 2.302585 3.135494 3.850148 4.032830 4.738382 6.408518
##        0 1.386294 2.197225 2.847503 3.433987 3.413786 3.951244 4.634729
##        1 3.135494 4.071254 4.738382 5.249621 5.471454 5.968668 7.408120
##      Max.        SD NAs
##  8.674026 1.2864326   0
##  7.826044 0.7800807   0
##  8.674026 1.0616036   0
## 
## Bootstrap summary: 
##           Variable Min.   5% 1st Qu. Median Mean 3rd Qu.  95% Max.   SD NAs
##   optimal_cutpoint 4.28 4.28    4.28   4.28 4.30    4.30 4.43 4.74 0.06   0
##              AUC_b 0.94 0.95    0.96   0.97 0.97    0.97 0.98 0.99 0.01   0
##            AUC_oob 0.92 0.95    0.96   0.97 0.96    0.97 0.98 0.99 0.01   0
##    sum_sens_spec_b 1.74 1.79    1.82   1.84 1.84    1.85 1.88 1.92 0.03   0
##  sum_sens_spec_oob 1.63 1.75    1.79   1.82 1.82    1.84 1.88 1.94 0.04   0
##              acc_b 0.86 0.89    0.90   0.91 0.91    0.92 0.93 0.95 0.01   0
##            acc_oob 0.84 0.87    0.89   0.90 0.90    0.91 0.93 0.96 0.02   0
##      sensitivity_b 0.80 0.90    0.93   0.94 0.94    0.96 0.97 0.99 0.02   0
##    sensitivity_oob 0.66 0.85    0.90   0.93 0.92    0.95 0.98 1.00 0.04   0
##      specificity_b 0.84 0.87    0.88   0.90 0.90    0.91 0.93 0.97 0.02   0
##    specificity_oob 0.82 0.85    0.88   0.89 0.89    0.91 0.93 0.97 0.02   0
##     cohens_kappa_b 0.69 0.75    0.77   0.79 0.79    0.81 0.84 0.89 0.03   0
##   cohens_kappa_oob 0.62 0.71    0.75   0.78 0.78    0.80 0.84 0.92 0.04   0
##              ppv_b 0.67 0.74    0.77   0.79 0.79    0.82 0.85 0.91 0.03   0
##            ppv_oob 0.65 0.72    0.76   0.79 0.79    0.81 0.86 0.93 0.04   0
##              npv_b 0.91 0.95    0.97   0.97 0.97    0.98 0.99 1.00 0.01   0
##            npv_oob 0.85 0.93    0.95   0.97 0.96    0.98 0.99 1.00 0.02   0

# ROC plot
plot(NEU.SCM)

# Youden index plot
plot_metric(NEU.SCM)

# %NEU SCM COW
PER.NEU.SCM <- cutpointr(Doff.by.cow.3, PERCENT.NEU.CO.COW, delaval.SCM.CO.COW, 
                            method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(PER.NEU.SCM)

## Method: maximize_metric 
## Predictor: PERCENT.NEU.CO.COW 
## Outcome: delaval.SCM.CO.COW 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC   n n_pos n_neg
##  0.5615 472   142   330
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv tp
##              60.7        1.2045 0.6907      0.3803      0.8242 0.4821 0.7556 54
##  fn fp  tn
##  88 58 272
## 
## Predictor summary: 
##     Data  Min.      5% 1st Qu. Median     Mean 3rd Qu.     95%   Max.       SD
##  Overall 15.69 29.2785 41.6925  51.14 50.71117 59.7675 70.3470 122.00 12.76777
##        0 15.69 30.7200 41.8025  50.00 49.87433 57.2675 68.6085  76.74 11.21753
##        1 19.90 27.4670 41.5350  54.27 52.65592 64.0825 73.3420 122.00 15.66793
##  NAs
##    0
##    0
##    0
## 
## Bootstrap summary: 
##           Variable  Min.    5% 1st Qu. Median  Mean 3rd Qu.   95%  Max.   SD
##   optimal_cutpoint 49.33 54.20   56.32  59.90 58.78   60.70 61.34 69.68 2.45
##              AUC_b  0.46  0.51    0.54   0.56  0.56    0.58  0.61  0.66 0.03
##            AUC_oob  0.36  0.49    0.53   0.56  0.56    0.59  0.63  0.71 0.04
##    sum_sens_spec_b  1.06  1.14    1.18   1.21  1.22    1.25  1.29  1.36 0.05
##  sum_sens_spec_oob  0.90  1.06    1.13   1.17  1.17    1.21  1.27  1.40 0.06
##              acc_b  0.57  0.63    0.66   0.68  0.68    0.70  0.72  0.76 0.03
##            acc_oob  0.43  0.59    0.63   0.66  0.66    0.69  0.72  0.77 0.04
##      sensitivity_b  0.14  0.33    0.38   0.42  0.43    0.48  0.54  0.68 0.07
##    sensitivity_oob  0.09  0.27    0.34   0.39  0.39    0.45  0.52  0.66 0.08
##      specificity_b  0.51  0.69    0.74   0.80  0.79    0.83  0.86  0.96 0.06
##    specificity_oob  0.40  0.66    0.73   0.79  0.77    0.83  0.86  0.96 0.07
##     cohens_kappa_b  0.06  0.15    0.19   0.22  0.22    0.25  0.30  0.36 0.05
##   cohens_kappa_oob -0.08  0.06    0.13   0.17  0.17    0.22  0.28  0.39 0.07
##              ppv_b  0.33  0.39    0.43   0.47  0.47    0.50  0.56  0.65 0.05
##            ppv_oob  0.18  0.33    0.39   0.43  0.43    0.48  0.55  0.69 0.07
##              npv_b  0.68  0.72    0.75   0.76  0.76    0.78  0.80  0.84 0.02
##            npv_oob  0.63  0.70    0.73   0.75  0.75    0.77  0.80  0.86 0.03
##  NAs
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0

# ROC plot
plot(PER.NEU.SCM)

# Youden index plot
plot_metric(PER.NEU.SCM)

# Lym
LYM.SCM  <- cutpointr(Doff.by.cow.3, log.TOTAL.LYM.CO.COW, delaval.SCM.CO.COW, 
                         method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(LYM.SCM)

## Method: maximize_metric 
## Predictor: log.TOTAL.LYM.CO.COW 
## Outcome: delaval.SCM.CO.COW 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC   n n_pos n_neg
##  0.9545 472   142   330
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv   npv  tp
##            3.3322        1.7904 0.8983      0.8873       0.903 0.7975 0.949 126
##  fn fp  tn
##  16 32 298
## 
## Predictor summary: 
##     Data     Min.       5%  1st Qu.   Median     Mean  3rd Qu.      95%
##  Overall 0.000000 1.098612 2.079442 2.833213 2.929250 3.663562 4.904507
##        0 0.000000 1.098612 1.945910 2.397895 2.381204 2.944439 3.555348
##        1 1.609438 3.044522 3.663562 4.077537 4.202878 4.632290 5.847005
##      Max.        SD NAs
##  7.082549 1.1672845   0
##  5.283204 0.7833538   0
##  7.082549 0.8849212   0
## 
## Bootstrap summary: 
##           Variable Min.   5% 1st Qu. Median Mean 3rd Qu.  95% Max.   SD NAs
##   optimal_cutpoint 2.94 3.18    3.33   3.34 3.36    3.40 3.56 3.64 0.11   0
##              AUC_b 0.91 0.94    0.95   0.95 0.95    0.96 0.97 0.98 0.01   0
##            AUC_oob 0.90 0.93    0.95   0.95 0.95    0.96 0.98 0.99 0.01   0
##    sum_sens_spec_b 1.69 1.75    1.78   1.80 1.80    1.82 1.85 1.89 0.03   0
##  sum_sens_spec_oob 1.54 1.69    1.73   1.77 1.76    1.79 1.84 1.89 0.05   0
##              acc_b 0.83 0.87    0.89   0.90 0.90    0.92 0.93 0.95 0.02   0
##            acc_oob 0.77 0.85    0.88   0.89 0.89    0.90 0.92 0.95 0.02   0
##      sensitivity_b 0.77 0.83    0.87   0.89 0.89    0.91 0.94 0.99 0.03   0
##    sensitivity_oob 0.61 0.77    0.83   0.87 0.86    0.90 0.94 1.00 0.05   0
##      specificity_b 0.78 0.85    0.89   0.91 0.91    0.93 0.95 0.98 0.03   0
##    specificity_oob 0.70 0.82    0.88   0.90 0.90    0.93 0.95 0.98 0.04   0
##     cohens_kappa_b 0.62 0.71    0.75   0.78 0.78    0.80 0.83 0.88 0.04   0
##   cohens_kappa_oob 0.52 0.66    0.71   0.75 0.74    0.77 0.82 0.89 0.05   0
##              ppv_b 0.61 0.72    0.78   0.81 0.81    0.84 0.89 0.94 0.05   0
##            ppv_oob 0.53 0.68    0.76   0.80 0.79    0.83 0.89 0.96 0.06   0
##              npv_b 0.90 0.93    0.94   0.95 0.95    0.96 0.97 1.00 0.01   0
##            npv_oob 0.85 0.90    0.93   0.94 0.94    0.95 0.97 1.00 0.02   0

# ROC plot
plot(LYM.SCM)

# Youden index plot
plot_metric(LYM.SCM)

# %LYM SCM COW
PER.LYM.SCM <- cutpointr(Doff.by.cow.3, PERCENT.LYM.CO.COW, delaval.SCM.CO.COW, 
                            method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 0

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(PER.LYM.SCM)

## Method: maximize_metric 
## Predictor: PERCENT.LYM.CO.COW 
## Outcome: delaval.SCM.CO.COW 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC   n n_pos n_neg
##  0.6581 472   330   142
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv
##              18.1        1.2454 0.5678      0.4848      0.7606 0.8247 0.3885
##   tp  fn fp  tn
##  160 170 34 108
## 
## Predictor summary: 
##     Data Min.     5% 1st Qu. Median     Mean 3rd Qu.     95% Max.       SD NAs
##  Overall 3.54 8.0955 12.7500 16.925 17.72085 21.2025 30.4885 69.0 7.392967   0
##        0 4.30 9.2030 13.7450 17.860 18.73639 22.9425 31.2500 54.5 7.233222   0
##        1 3.54 7.2440 11.0325 14.260 15.36077 18.0500 25.4925 69.0 7.243765   0
## 
## Bootstrap summary: 
##           Variable  Min.    5% 1st Qu. Median  Mean 3rd Qu.   95%  Max.   SD
##   optimal_cutpoint 10.09 13.01   14.04  17.19 16.59   18.10 19.13 24.40 2.13
##              AUC_b  0.56  0.61    0.64   0.66  0.66    0.68  0.70  0.74 0.03
##            AUC_oob  0.51  0.60    0.63   0.66  0.66    0.68  0.72  0.79 0.04
##    sum_sens_spec_b  1.12  1.20    1.24   1.27  1.27    1.30  1.34  1.41 0.04
##  sum_sens_spec_oob  1.00  1.09    1.15   1.19  1.19    1.23  1.29  1.42 0.06
##              acc_b  0.42  0.54    0.57   0.61  0.62    0.67  0.71  0.76 0.06
##            acc_oob  0.38  0.50    0.54   0.58  0.58    0.63  0.68  0.74 0.06
##      sensitivity_b  0.18  0.43    0.49   0.56  0.59    0.73  0.81  0.95 0.13
##    sensitivity_oob  0.19  0.38    0.46   0.53  0.57    0.70  0.80  0.94 0.14
##      specificity_b  0.23  0.45    0.55   0.72  0.68    0.79  0.84  0.96 0.13
##    specificity_oob  0.12  0.37    0.48   0.67  0.62    0.75  0.82  0.96 0.15
##     cohens_kappa_b  0.08  0.16    0.20   0.23  0.23    0.26  0.31  0.41 0.05
##   cohens_kappa_oob  0.00  0.07    0.12   0.16  0.16    0.20  0.25  0.37 0.05
##              ppv_b  0.72  0.76    0.79   0.82  0.82    0.84  0.88  0.95 0.04
##            ppv_oob  0.62  0.71    0.75   0.78  0.78    0.81  0.86  0.94 0.04
##              npv_b  0.31  0.36    0.39   0.42  0.43    0.46  0.52  0.67 0.05
##            npv_oob  0.23  0.31    0.35   0.39  0.39    0.42  0.48  0.60 0.05
##  NAs
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0

# ROC plot
plot(PER.LYM.SCM)

# Youden index plot
plot_metric(PER.LYM.SCM)

# MAC SCM COW
MAC.SCM <- cutpointr(Doff.by.cow.3, log.TOTAL.MAC.CO.COW, delaval.SCM.CO.COW, 
                        method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(MAC.SCM)

## Method: maximize_metric 
## Predictor: log.TOTAL.MAC.CO.COW 
## Outcome: delaval.SCM.CO.COW 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC   n n_pos n_neg
##  0.9457 472   142   330
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity   ppv    npv  tp
##            4.1744        1.7583 0.8983       0.831      0.9273 0.831 0.9273 118
##  fn fp  tn
##  24 24 306
## 
## Predictor summary: 
##     Data     Min.       5%  1st Qu.   Median     Mean  3rd Qu.      95%
##  Overall 0.000000 1.609438 2.564949 3.384247 3.486201 4.412744 5.738424
##        0 0.000000 1.098612 2.197225 2.944439 2.868476 3.526361 4.276666
##        1 2.772589 3.232757 4.412744 4.927254 4.921760 5.424950 6.410421
##      Max.        SD NAs
##  7.289611 1.3335800   0
##  6.601230 0.9724210   0
##  7.289611 0.8751049   0
## 
## Bootstrap summary: 
##           Variable Min.   5% 1st Qu. Median Mean 3rd Qu.  95% Max.   SD NAs
##   optimal_cutpoint 3.53 3.91    4.01   4.17 4.15    4.32 4.33 4.60 0.15   0
##              AUC_b 0.90 0.93    0.94   0.95 0.95    0.95 0.96 0.97 0.01   0
##            AUC_oob 0.89 0.92    0.94   0.95 0.95    0.96 0.97 0.99 0.01   0
##    sum_sens_spec_b 1.65 1.72    1.75   1.77 1.77    1.79 1.82 1.88 0.03   0
##  sum_sens_spec_oob 1.53 1.66    1.70   1.73 1.73    1.76 1.81 1.87 0.05   0
##              acc_b 0.83 0.87    0.89   0.90 0.90    0.91 0.93 0.95 0.02   0
##            acc_oob 0.75 0.84    0.87   0.89 0.88    0.90 0.92 0.95 0.02   0
##      sensitivity_b 0.71 0.78    0.82   0.85 0.85    0.88 0.91 0.96 0.04   0
##    sensitivity_oob 0.58 0.73    0.78   0.82 0.82    0.86 0.91 0.98 0.06   0
##      specificity_b 0.80 0.86    0.90   0.93 0.92    0.95 0.96 0.99 0.03   0
##    specificity_oob 0.69 0.84    0.88   0.92 0.91    0.94 0.97 1.00 0.04   0
##     cohens_kappa_b 0.61 0.70    0.74   0.77 0.76    0.79 0.82 0.87 0.04   0
##   cohens_kappa_oob 0.48 0.64    0.69   0.73 0.73    0.76 0.81 0.88 0.05   0
##              ppv_b 0.61 0.73    0.78   0.83 0.83    0.87 0.91 0.96 0.06   0
##            ppv_oob 0.52 0.68    0.75   0.81 0.81    0.86 0.92 1.00 0.07   0
##              npv_b 0.88 0.91    0.92   0.93 0.93    0.95 0.96 0.98 0.02   0
##            npv_oob 0.82 0.88    0.91   0.92 0.92    0.94 0.96 0.99 0.02   0

# ROC plot
plot(MAC.SCM)

# Youden index plot
plot_metric(MAC.SCM)

# %MAC SCM COW
PER.MAC.SCM <- cutpointr(Doff.by.cow.3, PERCENT.MAC.CO.COW, delaval.SCM.CO.COW, 
                            method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 0

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(PER.MAC.SCM)

## Method: maximize_metric 
## Predictor: PERCENT.MAC.CO.COW 
## Outcome: delaval.SCM.CO.COW 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##    AUC   n n_pos n_neg
##  0.492 472   330   142
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv
##             17.53        1.0819 0.6695      0.8636      0.2183 0.7197 0.4079
##   tp fn  fp tn
##  285 45 111 31
## 
## Predictor summary: 
##     Data Min.      5% 1st Qu. Median     Mean 3rd Qu.     95%  Max.       SD
##  Overall 3.41 12.3650 20.5825 29.845 31.87684   41.45 58.0275 74.51 14.27063
##        0 3.41 12.3350 21.1925 30.225 31.38961   40.72 52.9780 74.51 13.26877
##        1 5.85 13.2085 19.4250 29.545 33.00915   45.96 62.1030 72.70 16.35931
##  NAs
##    0
##    0
##    0
## 
## Bootstrap summary: 
##           Variable  Min.    5% 1st Qu. Median Mean 3rd Qu.   95% Max.   SD NAs
##   optimal_cutpoint  9.38 17.53   17.53  20.59  Inf   26.85 32.96  Inf  NaN   0
##              AUC_b  0.39  0.44    0.47   0.49 0.49    0.51  0.54 0.60 0.03   0
##            AUC_oob  0.36  0.43    0.46   0.49 0.49    0.52  0.56 0.65 0.04   0
##    sum_sens_spec_b  1.00  1.04    1.07   1.10 1.10    1.13  1.16 1.24 0.04   0
##  sum_sens_spec_oob  0.78  0.92    0.99   1.03 1.03    1.07  1.13 1.26 0.06   0
##              acc_b  0.28  0.51    0.58   0.65 0.63    0.67  0.70 0.75 0.06   0
##            acc_oob  0.24  0.46    0.54   0.61 0.60    0.66  0.70 0.74 0.08   0
##      sensitivity_b  0.00  0.46    0.63   0.80 0.74    0.86  0.89 0.98 0.15   0
##    sensitivity_oob  0.00  0.42    0.59   0.77 0.72    0.85  0.88 0.97 0.16   0
##      specificity_b  0.05  0.19    0.23   0.31 0.36    0.47  0.64 1.00 0.15   0
##    specificity_oob  0.00  0.15    0.21   0.27 0.31    0.40  0.57 1.00 0.14   0
##     cohens_kappa_b  0.00  0.04    0.07   0.10 0.10    0.13  0.17 0.24 0.04   0
##   cohens_kappa_oob -0.20 -0.07   -0.01   0.03 0.03    0.08  0.14 0.27 0.07   0
##              ppv_b  0.66  0.69    0.71   0.73 0.73    0.75  0.77 1.00 0.03   2
##            ppv_oob  0.25  0.65    0.69   0.71 0.71    0.73  0.76 0.81 0.04   7
##              npv_b  0.27  0.31    0.35   0.38 0.39    0.42  0.49 0.59 0.05   0
##            npv_oob  0.00  0.24    0.29   0.33 0.34    0.38  0.47 0.67 0.07   0

# ROC plot
plot(PER.MAC.SCM)

## Warning: Removed 2 rows containing non-finite outside the scale range
## (`stat_density()`).

# Youden index plot
plot_metric(PER.MAC.SCM)

# Summary of results when using SCM as reference

print(c("Optimal n NEU:", exp(NEU.SCM$optimal_cutpoint), NEU.SCM$sensitivity, NEU.SCM$specificity, NEU.SCM$ppv, NEU.SCM$npv, NEU.SCM$AUC))

##                                                 sensitivity         specificity 
##    "Optimal n NEU:"                "72" "0.943661971830986" "0.890909090909091" 
##                                                             
## "0.788235294117647" "0.973509933774834" "0.965215535638071"

print(c("Optimal %NEU:", PER.NEU.SCM$optimal_cutpoint, PER.NEU.SCM$sensitivity, PER.NEU.SCM$specificity, PER.NEU.SCM$ppv, PER.NEU.SCM$npv, PER.NEU.SCM$AUC))

##                                                 sensitivity         specificity 
##     "Optimal %NEU:"              "60.7" "0.380281690140845" "0.824242424242424" 
##                                                             
## "0.482142857142857" "0.755555555555556" "0.561491677336748"

print(c("Optimal n LYM:", exp(LYM.SCM$optimal_cutpoint), LYM.SCM$sensitivity, LYM.SCM$specificity, LYM.SCM$ppv, LYM.SCM$npv, LYM.SCM$AUC))

##                                                 sensitivity         specificity 
##    "Optimal n LYM:"                "28" "0.887323943661972" "0.903030303030303" 
##                                                             
##  "0.79746835443038" "0.949044585987261" "0.954492104139991"

print(c("Optimal %LYM:", PER.LYM.SCM$optimal_cutpoint, PER.LYM.SCM$sensitivity, PER.LYM.SCM$specificity, PER.LYM.SCM$ppv, PER.LYM.SCM$npv, PER.LYM.SCM$AUC))

##                                                 sensitivity         specificity 
##     "Optimal %LYM:"              "18.1" "0.484848484848485"  "0.76056338028169" 
##                                                             
## "0.824742268041237" "0.388489208633094" "0.658087921468203"

print(c("Optimal n MAC:", exp(MAC.SCM$optimal_cutpoint), MAC.SCM$sensitivity, MAC.SCM$specificity, MAC.SCM$ppv, MAC.SCM$npv, MAC.SCM$AUC))

##                                                 sensitivity         specificity 
##    "Optimal n MAC:"                "65" "0.830985915492958" "0.927272727272727" 
##                                                             
## "0.830985915492958" "0.927272727272727" "0.945657276995305"

print(c("Optimal %MAC:", PER.MAC.SCM$optimal_cutpoint, PER.MAC.SCM$sensitivity, PER.MAC.SCM$specificity, PER.MAC.SCM$ppv, PER.MAC.SCM$npv, PER.MAC.SCM$AUC))

##                                                 sensitivity         specificity 
##     "Optimal %MAC:"             "17.53" "0.863636363636364"  "0.21830985915493" 
##                                                             
##  "0.71969696969697" "0.407894736842105" "0.491954758856167"

# CI using bp

# Number neu
NEU.SCM_optimal_cutpoint <- boot_ci(NEU.SCM, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
NEU.SCM_lower_ci <- subset(NEU.SCM_optimal_cutpoint, NEU.SCM_optimal_cutpoint$quantile == 0.025)$values
exp(NEU.SCM_lower_ci)

## [1] 72

NEU.SCM_upper_ci <- subset(NEU.SCM_optimal_cutpoint, NEU.SCM_optimal_cutpoint$quantile == 0.975)$values
exp(NEU.SCM_upper_ci)

## [1] 86

boot_ci(NEU.SCM, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.892
## 2    0.975  0.978

boot_ci(NEU.SCM, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.860
## 2    0.975  0.931

boot_ci(NEU.SCM, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.728
## 2    0.975  0.862

boot_ci(NEU.SCM, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.951
## 2    0.975  0.990

boot_ci(NEU.SCM, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.949
## 2    0.975  0.980

# Percent neu
PER.NEU.SCM_optimal_cutpoint <- boot_ci(PER.NEU.SCM, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
PER.NEU.SCM_optimal_cutpoint

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025   54.1
## 2    0.975   62.4

PER.NEU.SCM_lower_ci <- subset(PER.NEU.SCM_optimal_cutpoint, PER.NEU.SCM_optimal_cutpoint$quantile == 0.025)$values
PER.NEU.SCM_upper_ci <- subset(PER.NEU.SCM_optimal_cutpoint, PER.NEU.SCM_optimal_cutpoint$quantile == 0.975)$values

boot_ci(PER.NEU.SCM, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.303
## 2    0.975  0.560

boot_ci(PER.NEU.SCM, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.665
## 2    0.975  0.872

boot_ci(PER.NEU.SCM, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.374
## 2    0.975  0.575

boot_ci(PER.NEU.SCM, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.716
## 2    0.975  0.807

boot_ci(PER.NEU.SCM, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.503
## 2    0.975  0.623

# Number lym
LYM.SCM_optimal_cutpoint <- boot_ci(LYM.SCM, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
LYM.SCM_lower_ci <- subset(LYM.SCM_optimal_cutpoint, LYM.SCM_optimal_cutpoint$quantile == 0.025)$values
exp(LYM.SCM_lower_ci)

## [1] 21

LYM.SCM_upper_ci <- subset(LYM.SCM_optimal_cutpoint, LYM.SCM_optimal_cutpoint$quantile == 0.975)$values
exp(LYM.SCM_upper_ci)

## [1] 35

boot_ci(LYM.SCM, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.824
## 2    0.975  0.953

boot_ci(LYM.SCM, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.828
## 2    0.975  0.958

boot_ci(LYM.SCM, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.696
## 2    0.975  0.897

boot_ci(LYM.SCM, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.922
## 2    0.975  0.978

boot_ci(LYM.SCM, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.932
## 2    0.975  0.973

# Percent lym
PER.LYM.SCM_optimal_cutpoint <- boot_ci(PER.LYM.SCM, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
PER.LYM.SCM_optimal_cutpoint

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025   13.0
## 2    0.975   19.2

PER.LYM.SCM_lower_ci <- subset(PER.LYM.SCM_optimal_cutpoint, PER.LYM.SCM_optimal_cutpoint$quantile == 0.025)$values
PER.LYM.SCM_upper_ci <- subset(PER.LYM.SCM_optimal_cutpoint, PER.LYM.SCM_optimal_cutpoint$quantile == 0.975)$values

boot_ci(PER.LYM.SCM, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.404
## 2    0.975  0.821

boot_ci(PER.LYM.SCM, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.427
## 2    0.975  0.861

boot_ci(PER.LYM.SCM, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.75 
## 2    0.975  0.888

boot_ci(PER.LYM.SCM, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.347
## 2    0.975  0.543

boot_ci(PER.LYM.SCM, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.604
## 2    0.975  0.708

# Number mac
MAC.SCM_optimal_cutpoint <- boot_ci(MAC.SCM, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
MAC.SCM_lower_ci <- subset(MAC.SCM_optimal_cutpoint, MAC.SCM_optimal_cutpoint$quantile == 0.025)$values
exp(MAC.SCM_lower_ci)

## [1] 45

MAC.SCM_upper_ci <- subset(MAC.SCM_optimal_cutpoint, MAC.SCM_optimal_cutpoint$quantile == 0.975)$values
exp(MAC.SCM_upper_ci)

## [1] 78

boot_ci(MAC.SCM, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.775
## 2    0.975  0.924

boot_ci(MAC.SCM, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.853
## 2    0.975  0.970

boot_ci(MAC.SCM, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.717
## 2    0.975  0.922

boot_ci(MAC.SCM, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.903
## 2    0.975  0.965

boot_ci(MAC.SCM, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.924
## 2    0.975  0.965

# Percent mac
PER.MAC.SCM_optimal_cutpoint <- boot_ci(PER.MAC.SCM, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
PER.MAC.SCM_optimal_cutpoint

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025   17.5
## 2    0.975   33.3

PER.MAC.SCM_lower_ci <- subset(PER.MAC.SCM_optimal_cutpoint, PER.MAC.SCM_optimal_cutpoint$quantile == 0.025)$values
PER.MAC.SCM_upper_ci <- subset(PER.MAC.SCM_optimal_cutpoint, PER.MAC.SCM_optimal_cutpoint$quantile == 0.975)$values

boot_ci(PER.MAC.SCM, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.440
## 2    0.975  0.897

boot_ci(PER.MAC.SCM, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.172
## 2    0.975  0.660

boot_ci(PER.MAC.SCM, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.682
## 2    0.975  0.781

boot_ci(PER.MAC.SCM, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.302
## 2    0.975  0.506

boot_ci(PER.MAC.SCM, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.435
## 2    0.975  0.552

# Prevalence

# True prevalence

table(Doff.by.cow.3$delaval.SCM.CO.COW)

## 
##   0   1 
## 341 145

prop.table(table(Doff.by.cow.3$delaval.SCM.CO.COW))

## 
##         0         1 
## 0.7016461 0.2983539

#Apparent prevalence

# ---- Neutrophils ----

# optimal cut-off

optimal_neu_doff_SCM_cow <- NEU.SCM$optimal_cutpoint

table(Doff.by.cow.3$TOTAL.NEU.CO.COW > exp(optimal_neu_doff_SCM_cow))

## 
## FALSE  TRUE 
##   304   169

prop.table(table(Doff.by.cow.3$TOTAL.NEU.CO.COW > exp(optimal_neu_doff_SCM_cow)))

## 
##     FALSE      TRUE 
## 0.6427061 0.3572939

# confidence intervals

table(Doff.by.cow.3$TOTAL.NEU.CO.COW > exp(NEU.SCM_lower_ci))

## 
## FALSE  TRUE 
##   304   169

prop.table(table(Doff.by.cow.3$TOTAL.NEU.CO.COW > exp(NEU.SCM_lower_ci)))

## 
##     FALSE      TRUE 
## 0.6427061 0.3572939

table(Doff.by.cow.3$TOTAL.NEU.CO.COW > exp(NEU.SCM_upper_ci))

## 
## FALSE  TRUE 
##   319   154

prop.table(table(Doff.by.cow.3$TOTAL.NEU.CO.COW >exp(NEU.SCM_upper_ci)))

## 
##     FALSE      TRUE 
## 0.6744186 0.3255814

# ---- Percent neu ----

# Optimal cut-off

optimal_per_neu_doff_SCM_cow <-PER.NEU.SCM$optimal_cutpoint

table(Doff.by.cow.3$PERCENT.NEU.CO.COW > optimal_per_neu_doff_SCM_cow)

## 
## FALSE  TRUE 
##   362   111

prop.table(table(Doff.by.cow.3$PERCENT.NEU.CO.COW > optimal_per_neu_doff_SCM_cow))

## 
##     FALSE      TRUE 
## 0.7653277 0.2346723

# confidence intervals

table(Doff.by.cow.3$PERCENT.NEU.CO.COW > PER.NEU.SCM_lower_ci)

## 
## FALSE  TRUE 
##   287   186

prop.table(table(Doff.by.cow.3$PERCENT.NEU.CO.COW > PER.NEU.SCM_lower_ci))

## 
##     FALSE      TRUE 
## 0.6067653 0.3932347

table(Doff.by.cow.3$PERCENT.NEU.CO.COW > PER.NEU.SCM_upper_ci)

## 
## FALSE  TRUE 
##   380    93

prop.table(table(Doff.by.cow.3$PERCENT.NEU.CO.COW >PER.NEU.SCM_upper_ci))

## 
##     FALSE      TRUE 
## 0.8033827 0.1966173

# ---- Lymphocytes ----

# optimal cut-off

optimal_lym_doff_SCM_cow <-LYM.SCM$optimal_cutpoint

table(Doff.by.cow.3$TOTAL.LYM.CO.COW > exp(optimal_lym_doff_SCM_cow))

## 
## FALSE  TRUE 
##   315   158

prop.table(table(Doff.by.cow.3$TOTAL.LYM.CO.COW > exp(optimal_lym_doff_SCM_cow)))

## 
##     FALSE      TRUE 
## 0.6659619 0.3340381

# confidence intervals

table(Doff.by.cow.3$TOTAL.LYM.CO.COW > exp(LYM.SCM_lower_ci))

## 
## FALSE  TRUE 
##   278   195

prop.table(table(Doff.by.cow.3$TOTAL.LYM.CO.COW > exp(LYM.SCM_lower_ci)))

## 
##     FALSE      TRUE 
## 0.5877378 0.4122622

table(Doff.by.cow.3$TOTAL.LYM.CO.COW > exp(LYM.SCM_upper_ci))

## 
## FALSE  TRUE 
##   338   135

prop.table(table(Doff.by.cow.3$TOTAL.LYM.CO.COW >exp(LYM.SCM_upper_ci)))

## 
##     FALSE      TRUE 
## 0.7145877 0.2854123

# ---- Percent lym ----

# optimal cut-off

optimal_per_lym_doff_SCM_cow <-PER.LYM.SCM$optimal_cutpoint

table(Doff.by.cow.3$PERCENT.LYM.CO.COW > optimal_per_lym_doff_SCM_cow)

## 
## FALSE  TRUE 
##   281   192

prop.table(table(Doff.by.cow.3$PERCENT.LYM.CO.COW > optimal_per_lym_doff_SCM_cow))

## 
##     FALSE      TRUE 
## 0.5940803 0.4059197

# confidence intervals

table(Doff.by.cow.3$PERCENT.LYM.CO.COW > PER.LYM.SCM_lower_ci)

## 
## FALSE  TRUE 
##   128   345

prop.table(table(Doff.by.cow.3$PERCENT.LYM.CO.COW > PER.LYM.SCM_lower_ci))

## 
##     FALSE      TRUE 
## 0.2706131 0.7293869

table(Doff.by.cow.3$PERCENT.LYM.CO.COW > PER.LYM.SCM_upper_ci)

## 
## FALSE  TRUE 
##   310   163

prop.table(table(Doff.by.cow.3$PERCENT.LYM.CO.COW >PER.LYM.SCM_upper_ci))

## 
##     FALSE      TRUE 
## 0.6553911 0.3446089

# ---- Macrophages ----

# optimal cut-off

optimal_mac_doff_SCM_cow <- MAC.SCM$optimal_cutpoint

table(Doff.by.cow.3$TOTAL.MAC.CO.COW > exp(optimal_mac_doff_SCM_cow))

## 
## FALSE  TRUE 
##   331   142

prop.table(table(Doff.by.cow.3$TOTAL.MAC.CO.COW > exp(optimal_mac_doff_SCM_cow)))

## 
##     FALSE      TRUE 
## 0.6997886 0.3002114

# confidence intervals

table(Doff.by.cow.3$TOTAL.MAC.CO.COW > exp(MAC.SCM_lower_ci))

## 
## FALSE  TRUE 
##   293   180

prop.table(table(Doff.by.cow.3$TOTAL.MAC.CO.COW > exp(MAC.SCM_lower_ci)))

## 
##     FALSE      TRUE 
## 0.6194503 0.3805497

table(Doff.by.cow.3$TOTAL.MAC.CO.COW > exp(MAC.SCM_upper_ci))

## 
## FALSE  TRUE 
##   350   123

prop.table(table(Doff.by.cow.3$TOTAL.MAC.CO.COW >exp(MAC.SCM_upper_ci)))

## 
##     FALSE      TRUE 
## 0.7399577 0.2600423

# Percent MAC

# optimal cut-off

optimal_per_mac_doff_SCM_cow <- PER.MAC.SCM$optimal_cutpoint

table(Doff.by.cow.3$PERCENT.MAC.CO.COW > optimal_per_mac_doff_SCM_cow)

## 
## FALSE  TRUE 
##    77   396

prop.table(table(Doff.by.cow.3$PERCENT.MAC.CO.COW > optimal_per_mac_doff_SCM_cow))

## 
##     FALSE      TRUE 
## 0.1627907 0.8372093

# confidence intervals

table(Doff.by.cow.3$PERCENT.MAC.CO.COW > PER.MAC.SCM_lower_ci)

## 
## FALSE  TRUE 
##    77   396

prop.table(table(Doff.by.cow.3$PERCENT.MAC.CO.COW > PER.MAC.SCM_lower_ci))

## 
##     FALSE      TRUE 
## 0.1627907 0.8372093

table(Doff.by.cow.3$PERCENT.MAC.CO.COW > PER.MAC.SCM_upper_ci)

## 
## FALSE  TRUE 
##   274   199

prop.table(table(Doff.by.cow.3$PERCENT.MAC.CO.COW >PER.MAC.SCM_upper_ci))

## 
##     FALSE      TRUE 
## 0.5792812 0.4207188

Quarter IMI

# ------ Quarter level -------


set.seed(123)

# SCC IMI quarter


SCC.IMI <- cutpointr(Doff.by.quarter.2, log.SCC, IMI, 
                        method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(SCC.IMI)

## Method: maximize_metric 
## Predictor: log.SCC 
## Outcome: IMI 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC    n n_pos n_neg
##  0.5876 1780   749  1031
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity   ppv    npv  tp
##            4.5539        1.1494 0.5888       0.486      0.6634 0.512 0.6399 364
##   fn  fp  tn
##  385 347 684
## 
## Predictor summary: 
##     Data     Min.       5%  1st Qu.   Median     Mean  3rd Qu.      95%
##  Overall 1.098612 2.564949 3.576476 4.290459 4.355470 5.057833 6.504438
##        0 1.098612 2.484907 3.465736 4.127134 4.191412 4.828314 5.978883
##        1 1.098612 2.639057 3.713572 4.499810 4.581296 5.365976 7.021595
##      Max.       SD NAs
##  8.589514 1.196636   0
##  8.239593 1.077550   0
##  8.589514 1.310959   0
## 
## Bootstrap summary: 
##           Variable Min.   5% 1st Qu. Median Mean 3rd Qu.  95% Max.   SD NAs
##   optimal_cutpoint 4.11 4.38    4.50   4.63 4.74    5.02 5.16 5.41 0.29   0
##              AUC_b 0.54 0.56    0.58   0.59 0.59    0.60 0.61 0.63 0.01   0
##            AUC_oob 0.53 0.56    0.58   0.59 0.59    0.60 0.62 0.66 0.02   0
##    sum_sens_spec_b 1.10 1.12    1.14   1.16 1.16    1.17 1.19 1.23 0.02   0
##  sum_sens_spec_oob 1.03 1.08    1.11   1.13 1.13    1.15 1.17 1.23 0.03   0
##              acc_b 0.54 0.58    0.59   0.60 0.60    0.61 0.63 0.65 0.02   0
##            acc_oob 0.52 0.55    0.57   0.59 0.59    0.60 0.62 0.65 0.02   0
##      sensitivity_b 0.23 0.30    0.36   0.46 0.44    0.52 0.56 0.64 0.09   0
##    sensitivity_oob 0.20 0.28    0.34   0.44 0.42    0.50 0.55 0.64 0.09   0
##      specificity_b 0.48 0.60    0.65   0.69 0.72    0.80 0.84 0.90 0.08   0
##    specificity_oob 0.45 0.58    0.63   0.68 0.70    0.79 0.84 0.89 0.09   0
##     cohens_kappa_b 0.10 0.13    0.15   0.16 0.16    0.18 0.20 0.24 0.02   0
##   cohens_kappa_oob 0.03 0.08    0.11   0.13 0.13    0.15 0.18 0.23 0.03   0
##              ppv_b 0.45 0.49    0.51   0.53 0.54    0.56 0.60 0.68 0.03   0
##            ppv_oob 0.41 0.46    0.49   0.51 0.52    0.54 0.59 0.67 0.04   0
##              npv_b 0.58 0.61    0.63   0.64 0.64    0.65 0.67 0.69 0.02   0
##            npv_oob 0.57 0.60    0.62   0.63 0.63    0.64 0.66 0.70 0.02   0

# ROC plot
plot(SCC.IMI)

# Youden index plot
plot_metric(SCC.IMI)

# NEU IMI COW
NEU.IMI <- cutpointr(Doff.by.quarter.2, log.Neu, IMI, 
                        method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(NEU.IMI)

## Method: maximize_metric 
## Predictor: log.Neu 
## Outcome: IMI 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC    n n_pos n_neg
##  0.5934 1771   745  1026
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv   npv  tp
##            4.0604        1.1598 0.6076      0.4054      0.7544 0.5451 0.636 302
##   fn  fp  tn
##  443 252 774
## 
## Predictor summary: 
##     Data      Min.       5%  1st Qu.   Median     Mean  3rd Qu.      95%
##  Overall 0.6931472 2.079442 2.772589 3.496508 3.678804 4.290459 6.008786
##        0 0.6931472 1.945910 2.708050 3.332205 3.487747 4.043051 5.394746
##        1 0.6931472 2.079442 2.944439 3.737670 3.941923 4.663439 6.859057
##      Max.       SD NAs
##  9.342946 1.266397   0
##  9.342946 1.100800   0
##  9.327056 1.423484   0
## 
## Bootstrap summary: 
##           Variable Min.   5% 1st Qu. Median Mean 3rd Qu.  95% Max.   SD NAs
##   optimal_cutpoint 3.18 3.69    3.87   4.06 4.01    4.16 4.22 4.83 0.19   0
##              AUC_b 0.55 0.57    0.58   0.59 0.59    0.60 0.62 0.64 0.01   0
##            AUC_oob 0.52 0.56    0.58   0.59 0.59    0.61 0.62 0.65 0.02   0
##    sum_sens_spec_b 1.10 1.13    1.15   1.17 1.17    1.18 1.20 1.24 0.02   0
##  sum_sens_spec_oob 1.01 1.09    1.12   1.14 1.14    1.16 1.19 1.24 0.03   0
##              acc_b 0.55 0.58    0.60   0.61 0.61    0.62 0.63 0.65 0.01   0
##            acc_oob 0.49 0.56    0.58   0.60 0.60    0.61 0.63 0.66 0.02   0
##      sensitivity_b 0.24 0.35    0.39   0.41 0.43    0.47 0.54 0.74 0.06   0
##    sensitivity_oob 0.16 0.34    0.37   0.40 0.42    0.45 0.52 0.67 0.06   0
##      specificity_b 0.43 0.63    0.70   0.76 0.74    0.78 0.81 0.91 0.06   0
##    specificity_oob 0.41 0.61    0.68   0.75 0.73    0.77 0.80 0.89 0.07   0
##     cohens_kappa_b 0.10 0.14    0.16   0.17 0.17    0.19 0.21 0.24 0.02   0
##   cohens_kappa_oob 0.01 0.09    0.12   0.15 0.15    0.17 0.20 0.26 0.03   0
##              ppv_b 0.46 0.50    0.53   0.55 0.55    0.57 0.59 0.67 0.03   0
##            ppv_oob 0.42 0.47    0.50   0.53 0.53    0.55 0.58 0.65 0.03   0
##              npv_b 0.60 0.62    0.63   0.64 0.64    0.65 0.67 0.70 0.02   0
##            npv_oob 0.57 0.60    0.62   0.63 0.63    0.64 0.66 0.69 0.02   0

# ROC plot
plot(NEU.IMI)

# Youden index plot
plot_metric(NEU.IMI)

# %NEU IMI quarter
PER.NEU.IMI <- cutpointr(Doff.by.quarter.2, PERCENT01_N, IMI, 
                            method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(PER.NEU.IMI)

## Method: maximize_metric 
## Predictor: PERCENT01_N 
## Outcome: IMI 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC    n n_pos n_neg
##  0.5159 1771   745  1026
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv
##             61.02        1.0497 0.5726      0.2242      0.8255 0.4827 0.5944
##   tp  fn  fp  tn
##  167 578 179 847
## 
## Predictor summary: 
##     Data  Min.     5% 1st Qu. Median     Mean 3rd Qu.     95%  Max.       SD
##  Overall 13.06 29.510 41.9600  50.00 50.30749  58.925 69.6450 85.71 12.11916
##        0 13.06 29.365 42.1025  50.00 50.04235  58.300 69.2225 85.71 11.88813
##        1 16.67 30.006 41.7900  50.74 50.67264  59.720 70.3340 84.62 12.42908
##  NAs
##    0
##    0
##    0
## 
## Bootstrap summary: 
##           Variable  Min.    5% 1st Qu. Median  Mean 3rd Qu.   95%  Max.   SD
##   optimal_cutpoint 25.86 48.50   51.52  57.30 56.31   60.70 61.30 71.72 5.29
##              AUC_b  0.47  0.49    0.51   0.52  0.52    0.53  0.54  0.57 0.01
##            AUC_oob  0.45  0.49    0.50   0.52  0.52    0.53  0.55  0.58 0.02
##    sum_sens_spec_b  1.00  1.03    1.05   1.06  1.06    1.07  1.10  1.14 0.02
##  sum_sens_spec_oob  0.92  0.97    1.00   1.02  1.02    1.04  1.07  1.12 0.03
##              acc_b  0.43  0.51    0.54   0.56  0.56    0.57  0.59  0.61 0.02
##            acc_oob  0.39  0.48    0.52   0.54  0.54    0.56  0.58  0.61 0.03
##      sensitivity_b  0.05  0.21    0.24   0.33  0.37    0.49  0.61  0.99 0.15
##    sensitivity_oob  0.02  0.18    0.23   0.31  0.34    0.45  0.58  0.98 0.14
##      specificity_b  0.02  0.46    0.57   0.73  0.70    0.82  0.84  0.98 0.15
##    specificity_oob  0.01  0.43    0.54   0.72  0.68    0.81  0.84  0.97 0.15
##     cohens_kappa_b  0.00  0.03    0.05   0.06  0.06    0.08  0.10  0.15 0.02
##   cohens_kappa_oob -0.08 -0.03    0.00   0.03  0.02    0.04  0.07  0.12 0.03
##              ppv_b  0.40  0.43    0.45   0.47  0.47    0.49  0.52  0.66 0.03
##            ppv_oob  0.30  0.39    0.42   0.44  0.44    0.46  0.50  0.57 0.03
##              npv_b  0.56  0.58    0.59   0.60  0.60    0.61  0.63  0.77 0.02
##            npv_oob  0.42  0.55    0.57   0.59  0.59    0.60  0.62  0.67 0.02
##  NAs
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0

# ROC plot
plot(PER.NEU.IMI)

# Youden index plot
plot_metric(PER.NEU.IMI)

# SCC IMI quarter
LYM.IMI <- cutpointr(Doff.by.quarter.2, log.Lym, IMI, method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(LYM.IMI)

## Method: maximize_metric 
## Predictor: log.Lym 
## Outcome: IMI 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC    n n_pos n_neg
##  0.5957 1771   745  1026
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv
##            2.8904        1.1722 0.6014      0.4899      0.6823 0.5282 0.6481
##   tp  fn  fp  tn
##  365 380 326 700
## 
## Predictor summary: 
##     Data Min.        5%  1st Qu.   Median     Mean  3rd Qu.      95%     Max.
##  Overall    0 0.6931472 1.791759 2.564949 2.646941 3.295837 4.803987 8.040769
##        0    0 0.6931472 1.791759 2.397895 2.467816 3.091042 4.427823 7.767687
##        1    0 1.0986123 1.945910 2.833213 2.893629 3.583519 5.388902 8.040769
##        SD NAs
##  1.227127   0
##  1.123296   0
##  1.318674   0
## 
## Bootstrap summary: 
##           Variable Min.   5% 1st Qu. Median Mean 3rd Qu.  95% Max.   SD NAs
##   optimal_cutpoint 2.30 2.77    2.77   2.89 2.88    2.89 3.09 3.43 0.12   0
##              AUC_b 0.54 0.57    0.59   0.60 0.60    0.60 0.62 0.64 0.01   0
##            AUC_oob 0.54 0.57    0.58   0.60 0.60    0.61 0.63 0.65 0.02   0
##    sum_sens_spec_b 1.09 1.14    1.16   1.18 1.18    1.19 1.21 1.24 0.02   0
##  sum_sens_spec_oob 1.05 1.11    1.14   1.16 1.16    1.18 1.21 1.25 0.03   0
##              acc_b 0.55 0.58    0.59   0.60 0.60    0.61 0.62 0.65 0.01   0
##            acc_oob 0.50 0.57    0.58   0.59 0.59    0.61 0.62 0.64 0.02   0
##      sensitivity_b 0.29 0.41    0.48   0.50 0.50    0.52 0.55 0.70 0.05   0
##    sensitivity_oob 0.27 0.38    0.47   0.49 0.49    0.52 0.55 0.70 0.05   0
##      specificity_b 0.45 0.63    0.65   0.68 0.68    0.70 0.76 0.85 0.04   0
##    specificity_oob 0.37 0.61    0.65   0.67 0.67    0.69 0.76 0.85 0.04   0
##     cohens_kappa_b 0.09 0.14    0.16   0.18 0.18    0.19 0.21 0.25 0.02   0
##   cohens_kappa_oob 0.05 0.11    0.14   0.16 0.16    0.18 0.21 0.26 0.03   0
##              ppv_b 0.45 0.50    0.52   0.53 0.53    0.54 0.57 0.64 0.02   0
##            ppv_oob 0.44 0.48    0.50   0.52 0.52    0.54 0.56 0.63 0.03   0
##              npv_b 0.58 0.62    0.64   0.65 0.65    0.66 0.68 0.69 0.02   0
##            npv_oob 0.58 0.61    0.63   0.64 0.64    0.66 0.68 0.71 0.02   0

# ROC plot
plot(LYM.IMI)

# Youden index plot
plot_metric(LYM.IMI)

# %LYM IMI quarter
PER.LYM.IMI <- cutpointr(Doff.by.quarter.2, PERCENT02_N, IMI, 
method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 0

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(PER.LYM.IMI)

## Method: maximize_metric 
## Predictor: PERCENT02_N 
## Outcome: IMI 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC    n n_pos n_neg
##  0.5173 1771  1026   745
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv
##             18.18         1.049 0.5206         0.5       0.549 0.6042 0.4436
##   tp  fn  fp  tn
##  513 513 336 409
## 
## Predictor summary: 
##     Data Min.    5% 1st Qu. Median     Mean 3rd Qu.    95%  Max.       SD NAs
##  Overall    0 7.805  13.325 17.760 19.04679 23.7650 34.360 60.00 8.224778   0
##        0    0 7.700  13.330 18.135 19.18767 24.1175 34.330 56.41 8.244352   0
##        1    0 8.052  13.250 17.380 18.85278 23.2900 34.374 60.00 8.199312   0
## 
## Bootstrap summary: 
##           Variable  Min.    5% 1st Qu. Median  Mean 3rd Qu.   95%  Max.   SD
##   optimal_cutpoint  6.17 15.79   16.10  18.18 18.70   19.51 26.54 45.00 3.32
##              AUC_b  0.47  0.49    0.51   0.52  0.52    0.53  0.54  0.58 0.01
##            AUC_oob  0.45  0.49    0.51   0.52  0.52    0.53  0.55  0.58 0.02
##    sum_sens_spec_b  1.00  1.03    1.05   1.06  1.06    1.08  1.10  1.14 0.02
##  sum_sens_spec_oob  0.92  0.97    1.00   1.02  1.02    1.04  1.07  1.11 0.03
##              acc_b  0.42  0.47    0.51   0.53  0.53    0.54  0.56  0.59 0.03
##            acc_oob  0.41  0.45    0.49   0.51  0.51    0.53  0.54  0.58 0.03
##      sensitivity_b  0.01  0.19    0.45   0.51  0.50    0.61  0.65  0.98 0.14
##    sensitivity_oob  0.01  0.17    0.43   0.49  0.48    0.59  0.64  0.97 0.14
##      specificity_b  0.02  0.41    0.45   0.56  0.56    0.61  0.85  0.99 0.13
##    specificity_oob  0.01  0.37    0.43   0.54  0.54    0.59  0.83  0.99 0.14
##     cohens_kappa_b  0.00  0.03    0.04   0.06  0.06    0.07  0.10  0.13 0.02
##   cohens_kappa_oob -0.08 -0.03    0.00   0.02  0.02    0.04  0.07  0.11 0.03
##              ppv_b  0.56  0.59    0.60   0.61  0.61    0.62  0.65  0.70 0.02
##            ppv_oob  0.46  0.55    0.58   0.59  0.59    0.61  0.63  0.71 0.02
##              npv_b  0.40  0.42    0.44   0.45  0.45    0.46  0.48  0.52 0.02
##            npv_oob  0.20  0.39    0.42   0.43  0.43    0.45  0.47  0.51 0.02
##  NAs
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0

# ROC plot
plot(PER.LYM.IMI)

# Youden index plot
plot_metric(PER.LYM.IMI)

# MAC IMI quarter
MAC.IMI <- cutpointr(Doff.by.quarter.2, log.Mac, IMI, method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(MAC.IMI)

## Method: maximize_metric 
## Predictor: log.Mac 
## Outcome: IMI 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC    n n_pos n_neg
##  0.5804 1771   745  1026
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv
##            3.9318        1.1242 0.5991      0.3289      0.7953 0.5385 0.6201
##   tp  fn  fp  tn
##  245 500 210 816
## 
## Predictor summary: 
##     Data Min.       5%  1st Qu.   Median     Mean  3rd Qu.      95%     Max.
##  Overall    0 1.098612 2.079442 2.944439 3.082324 3.970292 5.603920 8.582232
##        0    0 1.098612 1.945910 2.772589 2.902256 3.707399 5.199736 8.582232
##        1    0 1.098612 2.197225 3.135494 3.330311 4.317488 6.077632 8.423322
##        SD NAs
##  1.431457   0
##  1.329270   0
##  1.527805   0
## 
## Bootstrap summary: 
##           Variable  Min.   5% 1st Qu. Median Mean 3rd Qu.  95% Max.   SD NAs
##   optimal_cutpoint  2.40 2.48    2.83   3.26 3.33    3.89 4.01 4.84 0.57   0
##              AUC_b  0.54 0.56    0.57   0.58 0.58    0.59 0.60 0.63 0.01   0
##            AUC_oob  0.52 0.55    0.57   0.58 0.58    0.59 0.61 0.64 0.02   0
##    sum_sens_spec_b  1.07 1.10    1.12   1.14 1.14    1.15 1.17 1.22 0.02   0
##  sum_sens_spec_oob  0.99 1.05    1.08   1.10 1.10    1.12 1.15 1.18 0.03   0
##              acc_b  0.51 0.54    0.56   0.58 0.58    0.60 0.62 0.64 0.03   0
##            acc_oob  0.47 0.52    0.54   0.56 0.56    0.59 0.61 0.63 0.03   0
##      sensitivity_b  0.19 0.31    0.35   0.49 0.49    0.62 0.72 0.77 0.15   0
##    sensitivity_oob  0.12 0.29    0.33   0.45 0.47    0.60 0.70 0.76 0.15   0
##      specificity_b  0.35 0.41    0.52   0.65 0.65    0.79 0.82 0.92 0.15   0
##    specificity_oob  0.32 0.40    0.50   0.63 0.63    0.77 0.81 0.93 0.15   0
##     cohens_kappa_b  0.07 0.10    0.12   0.14 0.14    0.15 0.17 0.21 0.02   0
##   cohens_kappa_oob -0.01 0.05    0.08   0.10 0.10    0.12 0.15 0.19 0.03   0
##              ppv_b  0.43 0.46    0.48   0.51 0.51    0.54 0.57 0.66 0.04   0
##            ppv_oob  0.40 0.44    0.46   0.48 0.49    0.52 0.56 0.62 0.04   0
##              npv_b  0.58 0.61    0.62   0.64 0.64    0.65 0.68 0.71 0.02   0
##            npv_oob  0.55 0.59    0.61   0.62 0.62    0.64 0.66 0.70 0.02   0

# ROC plot
plot(MAC.IMI)

# Youden index plot
plot_metric(MAC.IMI)

# %MAC IMI COW
PER.MAC.IMI <- cutpointr(Doff.by.quarter.2, PERCENT03_N, IMI, method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 0

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(PER.MAC.IMI)

## Method: maximize_metric 
## Predictor: PERCENT03_N 
## Outcome: IMI 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC    n n_pos n_neg
##  0.5085 1771  1026   745
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv
##              25.3        1.0483 0.5381      0.6121      0.4362 0.5992 0.4495
##   tp  fn  fp  tn
##  628 398 420 325
## 
## Predictor summary: 
##     Data Min.    5% 1st Qu. Median     Mean 3rd Qu.    95%  Max.       SD NAs
##  Overall 0.00 9.720  19.355 29.350 30.64550 40.0000 57.615 83.54 14.58798   0
##        0 0.00 9.760  19.570 29.605 30.76740 39.8025 57.470 83.54 14.50642   0
##        1 1.64 9.728  19.120 28.800 30.47762 40.8000 57.582 73.41 14.70768   0
## 
## Bootstrap summary: 
##           Variable  Min.    5% 1st Qu. Median  Mean 3rd Qu.   95%  Max.   SD
##   optimal_cutpoint  2.94 21.21   21.82  25.30 25.26   26.59 33.01 74.37 5.54
##              AUC_b  0.46  0.49    0.50   0.51  0.51    0.52  0.53  0.56 0.01
##            AUC_oob  0.45  0.48    0.50   0.51  0.51    0.52  0.54  0.57 0.02
##    sum_sens_spec_b  1.00  1.02    1.04   1.06  1.06    1.07  1.10  1.13 0.02
##  sum_sens_spec_oob  0.91  0.97    1.01   1.03  1.03    1.05  1.08  1.13 0.03
##              acc_b  0.42  0.51    0.54   0.55  0.54    0.56  0.57  0.60 0.02
##            acc_oob  0.39  0.48    0.52   0.53  0.53    0.54  0.56  0.59 0.03
##      sensitivity_b  0.00  0.43    0.59   0.62  0.62    0.70  0.73  0.99 0.11
##    sensitivity_oob  0.00  0.40    0.57   0.61  0.61    0.69  0.73  1.00 0.11
##      specificity_b  0.02  0.32    0.36   0.44  0.43    0.47  0.62  1.00 0.11
##    specificity_oob  0.00  0.30    0.34   0.42  0.41    0.46  0.58  1.00 0.11
##     cohens_kappa_b  0.00  0.02    0.04   0.06  0.06    0.07  0.10  0.14 0.02
##   cohens_kappa_oob -0.09 -0.03    0.01   0.03  0.03    0.05  0.08  0.13 0.03
##              ppv_b  0.56  0.58    0.59   0.60  0.61    0.61  0.63  1.00 0.03
##            ppv_oob  0.38  0.55    0.58   0.59  0.59    0.60  0.62  1.00 0.03
##              npv_b  0.40  0.43    0.44   0.46  0.46    0.47  0.49  0.67 0.02
##            npv_oob  0.26  0.39    0.42   0.44  0.44    0.46  0.48  0.53 0.03
##  NAs
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    1

# ROC plot
plot(PER.MAC.IMI)

# Youden index plot
plot_metric(PER.MAC.IMI)

# Summary of results when using IMI as reference

print(c("Optimal SCC:", exp(SCC.IMI$optimal_cutpoint), SCC.IMI$sensitivity, SCC.IMI$specificity, SCC.IMI$ppv, SCC.IMI$npv, SCC.IMI$AUC))

##                                                 sensitivity         specificity 
##      "Optimal SCC:"                "95" "0.485981308411215" "0.663433559650824" 
##                                                             
## "0.511954992967651" "0.639850327408793" "0.587602739637331"

print(c("Optimal n NEU:", exp(NEU.IMI$optimal_cutpoint), NEU.IMI$sensitivity, NEU.IMI$specificity, NEU.IMI$ppv, NEU.IMI$npv, NEU.IMI$AUC))

##                                                 sensitivity         specificity 
##    "Optimal n NEU:"                "58" "0.405369127516779" "0.754385964912281" 
##                                                             
## "0.545126353790614" "0.635990139687757" "0.593420071431375"

print(c("Optimal %NEU:", PER.NEU.IMI$optimal_cutpoint, PER.NEU.IMI$sensitivity, PER.NEU.IMI$specificity, PER.NEU.IMI$ppv, PER.NEU.IMI$npv, PER.NEU.IMI$AUC))

##                                                 sensitivity         specificity 
##     "Optimal %NEU:"             "61.02" "0.224161073825503" "0.825536062378168" 
##                                                             
## "0.482658959537572" "0.594385964912281" "0.515860120098905"

print(c("Optimal n LYM:", exp(LYM.IMI$optimal_cutpoint), LYM.IMI$sensitivity, LYM.IMI$specificity, LYM.IMI$ppv, LYM.IMI$npv, LYM.IMI$AUC))

##                                                 sensitivity         specificity 
##    "Optimal n LYM:"                "18"  "0.48993288590604" "0.682261208576998" 
##                                                             
##  "0.52821997105644" "0.648148148148148" "0.595650012428536"

print(c("Optimal %LYM:", PER.LYM.IMI$optimal_cutpoint, PER.LYM.IMI$sensitivity, PER.LYM.IMI$specificity, PER.LYM.IMI$ppv, PER.LYM.IMI$npv, PER.LYM.IMI$AUC))

##                                                 sensitivity         specificity 
##     "Optimal %LYM:"             "18.18"               "0.5" "0.548993288590604" 
##                                                             
## "0.604240282685512" "0.443600867678959" "0.517328649737692"

print(c("Optimal n MAC:", exp(MAC.IMI$optimal_cutpoint), MAC.IMI$sensitivity, MAC.IMI$specificity, MAC.IMI$ppv, MAC.IMI$npv, MAC.IMI$AUC))

##                                                 sensitivity         specificity 
##    "Optimal n MAC:"                "51" "0.328859060402685" "0.795321637426901" 
##                                                             
## "0.538461538461538" "0.620060790273556" "0.580438792731269"

print(c("Optimal %MAC:", PER.MAC.IMI$optimal_cutpoint, PER.MAC.IMI$sensitivity, PER.MAC.IMI$specificity, PER.MAC.IMI$ppv, PER.MAC.IMI$npv, PER.MAC.IMI$AUC))

##                                                 sensitivity         specificity 
##     "Optimal %MAC:"              "25.3" "0.612085769980507" "0.436241610738255" 
##                                                             
## "0.599236641221374" "0.449515905947441" "0.508510930570274"

# CI using bootstrapping

# SCC
SCC.IMI.quarter_optimal_cutpoint <- boot_ci(SCC.IMI, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
SCC.IMI.quarter_lower_ci <- subset(SCC.IMI.quarter_optimal_cutpoint, SCC.IMI.quarter_optimal_cutpoint$quantile == 0.025)$values
exp(SCC.IMI.quarter_lower_ci)

## [1] 79

SCC.IMI.quarter_upper_ci <- subset(SCC.IMI.quarter_optimal_cutpoint, SCC.IMI.quarter_optimal_cutpoint$quantile == 0.975)$values
exp(SCC.IMI.quarter_upper_ci)

## [1] 202

boot_ci(SCC.IMI, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.285
## 2    0.975  0.574

boot_ci(SCC.IMI, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.587
## 2    0.975  0.862

boot_ci(SCC.IMI, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.484
## 2    0.975  0.615

boot_ci(SCC.IMI, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.608
## 2    0.975  0.671

boot_ci(SCC.IMI, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.560
## 2    0.975  0.614

# Number neu

NEU.IMI.quarter_optimal_cutpoint <- boot_ci(NEU.IMI, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
NEU.IMI.quarter_lower_ci <- subset(NEU.IMI.quarter_optimal_cutpoint, NEU.IMI.quarter_optimal_cutpoint$quantile == 0.025)$values
exp(NEU.IMI.quarter_lower_ci)

## [1] 35

NEU.IMI.quarter_upper_ci <- subset(NEU.IMI.quarter_optimal_cutpoint, NEU.IMI.quarter_optimal_cutpoint$quantile == 0.975)$values
exp(NEU.IMI.quarter_upper_ci)

## [1] 69.02483

boot_ci(NEU.IMI, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.342
## 2    0.975  0.570

boot_ci(NEU.IMI, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.593
## 2    0.975  0.815

boot_ci(NEU.IMI, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.490
## 2    0.975  0.598

boot_ci(NEU.IMI, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.612
## 2    0.975  0.672

boot_ci(NEU.IMI, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.568
## 2    0.975  0.621

# Percent neu
PER.NEU.IMI.quarter_optimal_cutpoint <- boot_ci(PER.NEU.IMI, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
PER.NEU.IMI.quarter_optimal_cutpoint

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025   47.4
## 2    0.975   65.2

PER.NEU.IMI.quarter_lower_ci <- subset(PER.NEU.IMI.quarter_optimal_cutpoint, PER.NEU.IMI.quarter_optimal_cutpoint$quantile == 0.025)$values
PER.NEU.IMI.quarter_upper_ci <- subset(PER.NEU.IMI.quarter_optimal_cutpoint, PER.NEU.IMI.quarter_optimal_cutpoint$quantile == 0.975)$values

boot_ci(PER.NEU.IMI, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.146
## 2    0.975  0.637

boot_ci(PER.NEU.IMI, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.431
## 2    0.975  0.899

boot_ci(PER.NEU.IMI, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.426
## 2    0.975  0.535

boot_ci(PER.NEU.IMI, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.574
## 2    0.975  0.635

boot_ci(PER.NEU.IMI, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.489
## 2    0.975  0.544

# Number lym
LYM.IMI.quarter_optimal_cutpoint <- boot_ci(LYM.IMI, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
LYM.IMI.quarter_lower_ci <- subset(LYM.IMI.quarter_optimal_cutpoint, LYM.IMI.quarter_optimal_cutpoint$quantile == 0.025)$values
exp(LYM.IMI.quarter_lower_ci)

## [1] 16

LYM.IMI.quarter_upper_ci <- subset(LYM.IMI.quarter_optimal_cutpoint, LYM.IMI.quarter_optimal_cutpoint$quantile == 0.975)$values
exp(LYM.IMI.quarter_upper_ci)

## [1] 27

boot_ci(LYM.IMI, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.352
## 2    0.975  0.561

boot_ci(LYM.IMI, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.618
## 2    0.975  0.813

boot_ci(LYM.IMI, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.491
## 2    0.975  0.576

boot_ci(LYM.IMI, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.620
## 2    0.975  0.680

boot_ci(LYM.IMI, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.567
## 2    0.975  0.623

# Percent lym
PER.LYM.IMI.quarter_optimal_cutpoint <- boot_ci(PER.LYM.IMI, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
PER.LYM.IMI.quarter_optimal_cutpoint

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025   15.4
## 2    0.975   28

PER.LYM.IMI.quarter_lower_ci <- subset(PER.LYM.IMI.quarter_optimal_cutpoint, PER.LYM.IMI.quarter_optimal_cutpoint$quantile == 0.025)$values
PER.LYM.IMI.quarter_upper_ci <- subset(PER.LYM.IMI.quarter_optimal_cutpoint, PER.LYM.IMI.quarter_optimal_cutpoint$quantile == 0.975)$values

boot_ci(PER.LYM.IMI, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.160
## 2    0.975  0.668

boot_ci(PER.LYM.IMI, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.392
## 2    0.975  0.879

boot_ci(PER.LYM.IMI, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.580
## 2    0.975  0.659

boot_ci(PER.LYM.IMI, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.419
## 2    0.975  0.488

boot_ci(PER.LYM.IMI, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.490
## 2    0.975  0.544

# Number mac
MAC.IMI.quarter_optimal_cutpoint <- boot_ci(MAC.IMI, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
MAC.IMI.quarter_lower_ci <- subset(MAC.IMI.quarter_optimal_cutpoint, MAC.IMI.quarter_optimal_cutpoint$quantile == 0.025)$values
exp(MAC.IMI.quarter_lower_ci)

## [1] 12

MAC.IMI.quarter_upper_ci <- subset(MAC.IMI.quarter_optimal_cutpoint, MAC.IMI.quarter_optimal_cutpoint$quantile == 0.975)$values
exp(MAC.IMI.quarter_upper_ci)

## [1] 57

boot_ci(MAC.IMI, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.297
## 2    0.975  0.733

boot_ci(MAC.IMI, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.398
## 2    0.975  0.826

boot_ci(MAC.IMI, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.456
## 2    0.975  0.581

boot_ci(MAC.IMI, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.604
## 2    0.975  0.686

boot_ci(MAC.IMI, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.554
## 2    0.975  0.607

# Percent mac
PER.MAC.IMI.quarter_optimal_cutpoint <- boot_ci(PER.MAC.IMI, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
PER.MAC.IMI.quarter_optimal_cutpoint

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025   20.7
## 2    0.975   36.7

PER.MAC.IMI.quarter_lower_ci <- subset(PER.MAC.IMI.quarter_optimal_cutpoint, PER.MAC.IMI.quarter_optimal_cutpoint$quantile == 0.025)$values
PER.MAC.IMI.quarter_upper_ci <- subset(PER.MAC.IMI.quarter_optimal_cutpoint, PER.MAC.IMI.quarter_optimal_cutpoint$quantile == 0.975)$values

boot_ci(PER.MAC.IMI, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.339
## 2    0.975  0.742

boot_ci(PER.MAC.IMI, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.309
## 2    0.975  0.700

boot_ci(PER.MAC.IMI, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.576
## 2    0.975  0.638

boot_ci(PER.MAC.IMI, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.421
## 2    0.975  0.499

boot_ci(PER.MAC.IMI, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.481
## 2    0.975  0.536

# Prevalence

# True prevalence

table(Doff.by.quarter.2$IMI)

## 
##    0    1 
## 1034  750

prop.table(table(Doff.by.quarter.2$IMI))

## 
##         0         1 
## 0.5795964 0.4204036

#Apparent prevalence

# ---- SCC ----

# optimal cut-off

optimal_scc_doff_IMI_quarter <- SCC.IMI$optimal_cutpoint

table(Doff.by.quarter.2$test_result_n > exp(optimal_scc_doff_IMI_quarter))

## 
## FALSE  TRUE 
##  1145   756

prop.table(table(Doff.by.quarter.2$test_result_n > exp(optimal_scc_doff_IMI_quarter)))

## 
##     FALSE      TRUE 
## 0.6023146 0.3976854

# confidence intervals

table(Doff.by.quarter.2$test_result_n > exp(SCC.IMI.quarter_lower_ci))

## 
## FALSE  TRUE 
##  1015   886

prop.table(table(Doff.by.quarter.2$test_result_n > exp(SCC.IMI.quarter_lower_ci)))

## 
##     FALSE      TRUE 
## 0.5339295 0.4660705

table(Doff.by.quarter.2$test_result_n > exp(SCC.IMI.quarter_upper_ci))

## 
## FALSE  TRUE 
##  1533   368

prop.table(table(Doff.by.quarter.2$test_result_n >exp(SCC.IMI.quarter_upper_ci)))

## 
##     FALSE      TRUE 
## 0.8064177 0.1935823

# ---- Neutrophils ----

# optimal cut-off

optimal_neu_doff_IMI_quarter <- NEU.IMI$optimal_cutpoint

table(Doff.by.quarter.2$TOTAL01_N > exp(optimal_neu_doff_IMI_quarter))

## 
## FALSE  TRUE 
##  1291   599

prop.table(table(Doff.by.quarter.2$TOTAL01_N > exp(optimal_neu_doff_IMI_quarter)))

## 
##     FALSE      TRUE 
## 0.6830688 0.3169312

# confidence intervals

table(Doff.by.quarter.2$TOTAL01_N > exp(NEU.IMI.quarter_lower_ci))

## 
## FALSE  TRUE 
##   986   904

prop.table(table(Doff.by.quarter.2$TOTAL01_N > exp(NEU.IMI.quarter_lower_ci)))

## 
##     FALSE      TRUE 
## 0.5216931 0.4783069

table(Doff.by.quarter.2$TOTAL01_N > exp(NEU.IMI.quarter_upper_ci))

## 
## FALSE  TRUE 
##  1384   506

prop.table(table(Doff.by.quarter.2$TOTAL01_N >exp(NEU.IMI.quarter_upper_ci)))

## 
##     FALSE      TRUE 
## 0.7322751 0.2677249

# ---- Percent neu ----

# Optimal cut-off

optimal_per_neu_doff_IMI_quarter <- PER.NEU.IMI$optimal_cutpoint

table(Doff.by.quarter.2$PERCENT01_N > optimal_per_neu_doff_IMI_quarter)

## 
## FALSE  TRUE 
##  1524   366

prop.table(table(Doff.by.quarter.2$PERCENT01_N > optimal_per_neu_doff_IMI_quarter))

## 
##     FALSE      TRUE 
## 0.8063492 0.1936508

# confidence intervals

table(Doff.by.quarter.2$PERCENT01_N > PER.NEU.IMI.quarter_lower_ci)

## 
## FALSE  TRUE 
##   768  1122

prop.table(table(Doff.by.quarter.2$PERCENT01_N > PER.NEU.IMI.quarter_lower_ci))

## 
##     FALSE      TRUE 
## 0.4063492 0.5936508

table(Doff.by.quarter.2$PERCENT01_N > PER.NEU.IMI.quarter_upper_ci)

## 
## FALSE  TRUE 
##  1674   216

prop.table(table(Doff.by.quarter.2$PERCENT01_N >PER.NEU.IMI.quarter_upper_ci))

## 
##     FALSE      TRUE 
## 0.8857143 0.1142857

# ---- Lymphocytes ----

# optimal cut-off

optimal_lym_doff_IMI_quarter <-LYM.IMI$optimal_cutpoint

table(Doff.by.quarter.2$TOTAL02_N > exp(optimal_lym_doff_IMI_quarter))

## 
## FALSE  TRUE 
##  1145   745

prop.table(table(Doff.by.quarter.2$TOTAL02_N > exp(optimal_lym_doff_IMI_quarter)))

## 
##     FALSE      TRUE 
## 0.6058201 0.3941799

# confidence intervals

table(Doff.by.quarter.2$TOTAL02_N > exp(LYM.IMI.quarter_lower_ci))

## 
## FALSE  TRUE 
##  1067   823

prop.table(table(Doff.by.quarter.2$TOTAL02_N > exp(LYM.IMI.quarter_lower_ci)))

## 
##     FALSE      TRUE 
## 0.5645503 0.4354497

table(Doff.by.quarter.2$TOTAL02_N > exp(LYM.IMI.quarter_upper_ci))

## 
## FALSE  TRUE 
##  1417   473

prop.table(table(Doff.by.quarter.2$TOTAL02_N >exp(LYM.IMI.quarter_upper_ci)))

## 
##     FALSE      TRUE 
## 0.7497354 0.2502646

# ---- Percent lym ----

# optimal cut-off

optimal_per_lym_doff_IMI_quarter <-PER.LYM.IMI$optimal_cutpoint

table(Doff.by.quarter.2$PERCENT02_N > optimal_per_lym_doff_IMI_quarter)

## 
## FALSE  TRUE 
##   985   905

prop.table(table(Doff.by.quarter.2$PERCENT02_N > optimal_per_lym_doff_IMI_quarter))

## 
##    FALSE     TRUE 
## 0.521164 0.478836

# confidence intervals

table(Doff.by.quarter.2$PERCENT02_N > PER.LYM.IMI.quarter_lower_ci)

## 
## FALSE  TRUE 
##   667  1223

prop.table(table(Doff.by.quarter.2$PERCENT02_N > PER.LYM.IMI.quarter_lower_ci))

## 
##     FALSE      TRUE 
## 0.3529101 0.6470899

table(Doff.by.quarter.2$PERCENT02_N > PER.LYM.IMI.quarter_upper_ci)

## 
## FALSE  TRUE 
##  1626   264

prop.table(table(Doff.by.quarter.2$PERCENT02_N >PER.LYM.IMI.quarter_upper_ci))

## 
##     FALSE      TRUE 
## 0.8603175 0.1396825

# ---- Macrophages ----

# optimal cut-off

optimal_mac_doff_IMI_quarter <- MAC.IMI$optimal_cutpoint

table(Doff.by.quarter.2$TOTAL03_N > exp(optimal_mac_doff_IMI_quarter))

## 
## FALSE  TRUE 
##  1411   479

prop.table(table(Doff.by.quarter.2$TOTAL03_N > exp(optimal_mac_doff_IMI_quarter)))

## 
##     FALSE      TRUE 
## 0.7465608 0.2534392

# confidence intervals

table(Doff.by.quarter.2$TOTAL03_N > exp(MAC.IMI.quarter_lower_ci))

## 
## FALSE  TRUE 
##   721  1169

prop.table(table(Doff.by.quarter.2$TOTAL03_N > exp(MAC.IMI.quarter_lower_ci)))

## 
##     FALSE      TRUE 
## 0.3814815 0.6185185

table(Doff.by.quarter.2$TOTAL03_N > exp(MAC.IMI.quarter_upper_ci))

## 
## FALSE  TRUE 
##  1443   447

prop.table(table(Doff.by.quarter.2$TOTAL03_N >exp(MAC.IMI.quarter_upper_ci)))

## 
##     FALSE      TRUE 
## 0.7634921 0.2365079

# Percent MAC

# optimal cut-off

optimal_per_mac_doff_IMI_quarter <- PER.MAC.IMI$optimal_cutpoint

table(Doff.by.quarter.2$PERCENT03_N > optimal_per_mac_doff_IMI_quarter)

## 
## FALSE  TRUE 
##   779  1111

prop.table(table(Doff.by.quarter.2$PERCENT03_N > optimal_per_mac_doff_IMI_quarter))

## 
##     FALSE      TRUE 
## 0.4121693 0.5878307

# confidence intervals

table(Doff.by.quarter.2$PERCENT03_N > PER.MAC.IMI.quarter_lower_ci)

## 
## FALSE  TRUE 
##   544  1346

prop.table(table(Doff.by.quarter.2$PERCENT03_N > PER.MAC.IMI.quarter_lower_ci))

## 
##     FALSE      TRUE 
## 0.2878307 0.7121693

table(Doff.by.quarter.2$PERCENT03_N > PER.MAC.IMI.quarter_upper_ci)

## 
## FALSE  TRUE 
##  1277   613

prop.table(table(Doff.by.quarter.2$PERCENT03_N >PER.MAC.IMI.quarter_upper_ci))

## 
##     FALSE      TRUE 
## 0.6756614 0.3243386

Quarter high SCC

# Set a seed for reproducibility

set.seed(123)

# Qscout subclinical mastitis reference

# n NEU SCM quarter
NEU.delaval.SCM <- cutpointr(Doff.by.quarter.2, log.Neu, delaval.SCM, 
                                 method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(NEU.delaval.SCM)

## Method: maximize_metric 
## Predictor: log.Neu 
## Outcome: delaval.SCM 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC    n n_pos n_neg
##  0.9671 1886   371  1515
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv
##            4.3175        1.8387 0.9162      0.9245      0.9142 0.7252 0.9802
##   tp fn  fp   tn
##  343 28 130 1385
## 
## Predictor summary: 
##     Data      Min.       5%  1st Qu.   Median     Mean  3rd Qu.      95%
##  Overall 0.6931472 2.079442 2.772589 3.496508 3.685987 4.317488 6.012472
##        0 0.6931472 1.945910 2.673554 3.258097 3.238194 3.806662 4.532599
##        1 2.0794415 4.127134 4.803987 5.231109 5.514573 6.017375 7.957542
##      Max.        SD NAs
##  9.342946 1.2632571   0
##  7.330405 0.7930368   0
##  9.342946 1.1759344   0
## 
## Bootstrap summary: 
##           Variable Min.   5% 1st Qu. Median Mean 3rd Qu.  95% Max.   SD NAs
##   optimal_cutpoint 4.11 4.28    4.32   4.32 4.33    4.34 4.37 4.66 0.04   0
##              AUC_b 0.95 0.96    0.96   0.97 0.97    0.97 0.98 0.98 0.01   0
##            AUC_oob 0.94 0.95    0.96   0.97 0.97    0.97 0.98 0.99 0.01   0
##    sum_sens_spec_b 1.79 1.82    1.83   1.84 1.84    1.85 1.86 1.89 0.01   0
##  sum_sens_spec_oob 1.72 1.79    1.82   1.83 1.83    1.85 1.87 1.90 0.02   0
##              acc_b 0.87 0.90    0.91   0.92 0.92    0.92 0.93 0.95 0.01   0
##            acc_oob 0.86 0.90    0.91   0.92 0.91    0.92 0.93 0.95 0.01   0
##      sensitivity_b 0.86 0.90    0.92   0.92 0.92    0.93 0.95 0.97 0.01   0
##    sensitivity_oob 0.76 0.88    0.90   0.92 0.92    0.93 0.95 0.98 0.02   0
##      specificity_b 0.85 0.90    0.91   0.92 0.92    0.92 0.93 0.97 0.01   0
##    specificity_oob 0.84 0.89    0.91   0.92 0.91    0.92 0.93 0.96 0.01   0
##     cohens_kappa_b 0.66 0.73    0.75   0.76 0.76    0.78 0.79 0.85 0.02   0
##   cohens_kappa_oob 0.65 0.71    0.74   0.76 0.76    0.77 0.80 0.84 0.03   0
##              ppv_b 0.59 0.69    0.71   0.73 0.73    0.75 0.77 0.89 0.03   0
##            ppv_oob 0.60 0.67    0.71   0.73 0.73    0.75 0.78 0.85 0.03   0
##              npv_b 0.96 0.97    0.98   0.98 0.98    0.98 0.99 0.99 0.00   0
##            npv_oob 0.94 0.97    0.97   0.98 0.98    0.98 0.99 0.99 0.01   0

# ROC plot
plot(NEU.delaval.SCM)

# Youden index plot
plot_metric(NEU.delaval.SCM)

# %NEU SCM quarter
PER.NEU.delaval.SCM <- cutpointr(Doff.by.quarter.2, PERCENT01_N, delaval.SCM, 
                                 method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(PER.NEU.delaval.SCM)

## Method: maximize_metric 
## Predictor: PERCENT01_N 
## Outcome: delaval.SCM 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC    n n_pos n_neg
##  0.5144 1886   371  1515
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv
##              58.7        1.1107 0.6813      0.3477       0.763 0.2643 0.8269
##   tp  fn  fp   tn
##  129 242 359 1156
## 
## Predictor summary: 
##     Data  Min.      5% 1st Qu. Median     Mean 3rd Qu.     95%  Max.       SD
##  Overall 13.06 29.9925  41.980  50.00 50.35406  58.895 69.7675 85.71 12.06358
##        0 15.10 30.8570  42.410  50.00 50.28239  58.080 69.1300 85.71 11.45665
##        1 13.06 26.4800  39.155  50.89 50.64674  62.280 72.3100 83.51 14.29086
##  NAs
##    0
##    0
##    0
## 
## Bootstrap summary: 
##           Variable  Min.    5% 1st Qu. Median  Mean 3rd Qu.   95%  Max.   SD
##   optimal_cutpoint 53.45 57.74   58.70  58.71 60.33   62.62 64.17 70.86 2.51
##              AUC_b  0.46  0.48    0.50   0.51  0.51    0.53  0.54  0.57 0.02
##            AUC_oob  0.44  0.47    0.50   0.51  0.51    0.53  0.55  0.59 0.02
##    sum_sens_spec_b  1.03  1.08    1.10   1.12  1.12    1.13  1.16  1.20 0.02
##  sum_sens_spec_oob  0.95  1.04    1.07   1.09  1.09    1.11  1.14  1.22 0.03
##              acc_b  0.57  0.66    0.68   0.69  0.71    0.74  0.76  0.80 0.04
##            acc_oob  0.54  0.65    0.67   0.69  0.70    0.73  0.76  0.80 0.04
##      sensitivity_b  0.09  0.21    0.25   0.34  0.31    0.37  0.40  0.52 0.07
##    sensitivity_oob  0.05  0.18    0.23   0.31  0.29    0.35  0.38  0.46 0.07
##      specificity_b  0.58  0.74    0.76   0.77  0.80    0.86  0.89  0.97 0.06
##    specificity_oob  0.57  0.72    0.75   0.77  0.80    0.86  0.89  0.97 0.06
##     cohens_kappa_b  0.03  0.08    0.10   0.11  0.11    0.13  0.15  0.21 0.02
##   cohens_kappa_oob -0.03  0.03    0.07   0.09  0.09    0.11  0.14  0.22 0.03
##              ppv_b  0.21  0.24    0.26   0.28  0.29    0.31  0.35  0.43 0.03
##            ppv_oob  0.16  0.22    0.24   0.26  0.27    0.29  0.33  0.45 0.04
##              npv_b  0.79  0.81    0.82   0.83  0.83    0.83  0.84  0.86 0.01
##            npv_oob  0.78  0.80    0.81   0.82  0.82    0.83  0.84  0.86 0.01
##  NAs
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0

# ROC plot
plot(PER.NEU.delaval.SCM)

# Youden index plot
plot_metric(PER.NEU.delaval.SCM)

# LYM SCM quarter
LYM.delaval.SCM <- cutpointr(Doff.by.quarter.2, log.Lym, delaval.SCM, method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(LYM.delaval.SCM)

## Method: maximize_metric 
## Predictor: log.Lym 
## Outcome: delaval.SCM 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC    n n_pos n_neg
##  0.9554 1886   371  1515
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv
##            3.3322        1.7917 0.9014      0.8868       0.905 0.6956 0.9703
##   tp fn  fp   tn
##  329 42 144 1371
## 
## Predictor summary: 
##     Data      Min.        5%  1st Qu.   Median     Mean  3rd Qu.      95%
##  Overall 0.0000000 0.7945135 1.791759 2.564949 2.659278 3.332205 4.818257
##        0 0.0000000 0.6931472 1.609438 2.302585 2.250548 2.890372 3.555348
##        1 0.6931472 2.8903718 3.688879 4.248495 4.328348 4.835998 6.326047
##      Max.        SD NAs
##  8.040769 1.2220077   0
##  6.612041 0.8668058   0
##  8.040769 1.0275960   0
## 
## Bootstrap summary: 
##           Variable Min.   5% 1st Qu. Median Mean 3rd Qu.  95% Max.   SD NAs
##   optimal_cutpoint 3.09 3.22    3.26   3.33 3.31    3.33 3.43 3.66 0.06   0
##              AUC_b 0.93 0.94    0.95   0.96 0.96    0.96 0.97 0.97 0.01   0
##            AUC_oob 0.93 0.94    0.95   0.96 0.96    0.96 0.97 0.98 0.01   0
##    sum_sens_spec_b 1.73 1.77    1.78   1.79 1.79    1.81 1.82 1.85 0.02   0
##  sum_sens_spec_oob 1.68 1.74    1.76   1.78 1.78    1.80 1.82 1.87 0.03   0
##              acc_b 0.85 0.88    0.89   0.90 0.90    0.91 0.92 0.94 0.01   0
##            acc_oob 0.82 0.87    0.89   0.90 0.89    0.90 0.91 0.93 0.01   0
##      sensitivity_b 0.81 0.86    0.88   0.90 0.89    0.91 0.93 0.95 0.02   0
##    sensitivity_oob 0.72 0.83    0.87   0.89 0.89    0.91 0.93 0.96 0.03   0
##      specificity_b 0.83 0.87    0.89   0.90 0.90    0.91 0.92 0.97 0.02   0
##    specificity_oob 0.80 0.86    0.88   0.90 0.90    0.91 0.93 0.97 0.02   0
##     cohens_kappa_b 0.60 0.67    0.69   0.71 0.71    0.73 0.76 0.81 0.03   0
##   cohens_kappa_oob 0.55 0.65    0.68   0.70 0.70    0.72 0.75 0.79 0.03   0
##              ppv_b 0.56 0.63    0.66   0.69 0.69    0.71 0.75 0.85 0.04   0
##            ppv_oob 0.51 0.61    0.65   0.68 0.68    0.71 0.75 0.86 0.04   0
##              npv_b 0.95 0.96    0.97   0.97 0.97    0.98 0.98 0.99 0.01   0
##            npv_oob 0.93 0.96    0.97   0.97 0.97    0.98 0.98 0.99 0.01   0

# ROC plot
plot(LYM.delaval.SCM)

# Youden index plot
plot_metric(LYM.delaval.SCM)

# %LYM SCM quarter
PER.LYM.delaval.SCM <- cutpointr(Doff.by.quarter.2, PERCENT02_N, delaval.SCM, 
                                 method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 0

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(PER.LYM.delaval.SCM)

## Method: maximize_metric 
## Predictor: PERCENT02_N 
## Outcome: delaval.SCM 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC    n n_pos n_neg
##  0.6399 1886  1515   371
## 
##  optimal_cutpoint sum_sens_spec   acc sensitivity specificity    ppv    npv  tp
##             19.72        1.2379 0.517      0.4508      0.7871 0.8963 0.2598 683
##   fn fp  tn
##  832 79 292
## 
## Predictor summary: 
##     Data Min.     5% 1st Qu. Median     Mean 3rd Qu.    95%  Max.       SD NAs
##  Overall  0.0 7.8925  13.465  17.84 19.11998  23.810 34.370 60.00 8.135124   0
##        0  0.0 8.2970  14.005  18.60 19.89336  24.840 35.234 60.00 8.391717   0
##        1  2.8 7.2150  11.505  15.70 15.96186  19.275 26.115 48.54 6.046239   0
## 
## Bootstrap summary: 
##           Variable  Min.    5% 1st Qu. Median  Mean 3rd Qu.   95%  Max.   SD
##   optimal_cutpoint 12.70 19.21   19.72  19.83 20.21   20.69 21.67 23.44 0.91
##              AUC_b  0.58  0.62    0.63   0.64  0.64    0.65  0.66  0.69 0.01
##            AUC_oob  0.57  0.61    0.63   0.64  0.64    0.65  0.67  0.70 0.02
##    sum_sens_spec_b  1.16  1.21    1.23   1.24  1.24    1.26  1.28  1.33 0.02
##  sum_sens_spec_oob  1.04  1.17    1.20   1.22  1.22    1.24  1.27  1.35 0.03
##              acc_b  0.40  0.46    0.49   0.51  0.51    0.52  0.54  0.75 0.03
##            acc_oob  0.40  0.45    0.48   0.50  0.50    0.52  0.54  0.69 0.03
##      sensitivity_b  0.29  0.37    0.41   0.44  0.43    0.46  0.48  0.83 0.04
##    sensitivity_oob  0.27  0.36    0.40   0.43  0.43    0.45  0.48  0.81 0.04
##      specificity_b  0.40  0.75    0.79   0.81  0.81    0.84  0.87  0.93 0.04
##    specificity_oob  0.25  0.71    0.77   0.79  0.79    0.82  0.87  0.92 0.05
##     cohens_kappa_b  0.08  0.11    0.12   0.14  0.14    0.15  0.16  0.21 0.02
##   cohens_kappa_oob  0.03  0.09    0.11   0.12  0.12    0.13  0.15  0.20 0.02
##              ppv_b  0.85  0.88    0.89   0.90  0.90    0.91  0.93  0.95 0.01
##            ppv_oob  0.78  0.86    0.88   0.89  0.89    0.91  0.92  0.95 0.02
##              npv_b  0.22  0.24    0.25   0.26  0.26    0.27  0.28  0.35 0.01
##            npv_oob  0.20  0.23    0.24   0.25  0.25    0.27  0.28  0.31 0.02
##  NAs
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0

# ROC plot
plot(PER.LYM.delaval.SCM)

# Youden index plot
plot_metric(PER.LYM.delaval.SCM)

# MAC SCM quarter
MAC.delaval.SCM <- cutpointr(Doff.by.quarter.2, log.Mac, delaval.SCM, method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(MAC.delaval.SCM)

## Method: maximize_metric 
## Predictor: log.Mac 
## Outcome: delaval.SCM 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC    n n_pos n_neg
##  0.9469 1886   371  1515
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv
##             3.989        1.7769 0.8977      0.8733      0.9036 0.6894 0.9668
##   tp fn  fp   tn
##  324 47 146 1369
## 
## Predictor summary: 
##     Data     Min.       5%  1st Qu.   Median     Mean  3rd Qu.      95%
##  Overall 0.000000 1.098612 2.079442 2.944439 3.086039 3.970292 5.574891
##        0 0.000000 1.098612 1.945910 2.564949 2.615936 3.295837 4.343805
##        1 1.098612 3.314021 4.382027 4.962845 5.005731 5.585372 7.015762
##      Max.       SD NAs
##  8.582232 1.422124   0
##  7.606885 1.042871   0
##  8.582232 1.119449   0
## 
## Bootstrap summary: 
##           Variable Min.   5% 1st Qu. Median Mean 3rd Qu.  95% Max.   SD NAs
##   optimal_cutpoint 3.56 3.87    3.87   3.95 3.94    3.99 4.04 4.25 0.07   0
##              AUC_b 0.92 0.94    0.94   0.95 0.95    0.95 0.96 0.97 0.01   0
##            AUC_oob 0.91 0.93    0.94   0.95 0.95    0.95 0.96 0.98 0.01   0
##    sum_sens_spec_b 1.72 1.75    1.77   1.78 1.78    1.79 1.81 1.84 0.02   0
##  sum_sens_spec_oob 1.66 1.72    1.75   1.77 1.77    1.79 1.81 1.84 0.02   0
##              acc_b 0.84 0.88    0.89   0.89 0.89    0.90 0.91 0.93 0.01   0
##            acc_oob 0.83 0.87    0.88   0.89 0.89    0.90 0.91 0.92 0.01   0
##      sensitivity_b 0.82 0.85    0.87   0.89 0.89    0.90 0.92 0.95 0.02   0
##    sensitivity_oob 0.73 0.82    0.86   0.88 0.88    0.90 0.92 0.95 0.03   0
##      specificity_b 0.82 0.87    0.88   0.90 0.90    0.91 0.92 0.95 0.02   0
##    specificity_oob 0.80 0.86    0.88   0.89 0.89    0.91 0.92 0.94 0.02   0
##     cohens_kappa_b 0.60 0.66    0.68   0.70 0.70    0.72 0.74 0.78 0.03   0
##   cohens_kappa_oob 0.56 0.64    0.67   0.69 0.69    0.71 0.73 0.78 0.03   0
##              ppv_b 0.56 0.62    0.65   0.68 0.68    0.70 0.73 0.81 0.03   0
##            ppv_oob 0.52 0.60    0.64   0.67 0.67    0.69 0.73 0.78 0.04   0
##              npv_b 0.95 0.96    0.97   0.97 0.97    0.97 0.98 0.99 0.01   0
##            npv_oob 0.93 0.95    0.96   0.97 0.97    0.97 0.98 0.99 0.01   0

# ROC plot
plot(MAC.delaval.SCM)

# Youden index plot
plot_metric(MAC.delaval.SCM)

# %MAC SCM quarter
PER.MAC.delaval.SCM<- cutpointr(Doff.by.quarter.2, PERCENT03_N, delaval.SCM, method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(PER.MAC.delaval.SCM)

## Method: maximize_metric 
## Predictor: PERCENT03_N 
## Outcome: delaval.SCM 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##    AUC    n n_pos n_neg
##  0.553 1886   371  1515
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv tp
##             47.98        1.1133 0.7635      0.2156      0.8977 0.3404 0.8237 80
##   fn  fp   tn
##  291 155 1360
## 
## Predictor summary: 
##     Data Min.     5% 1st Qu. Median     Mean 3rd Qu.    95%  Max.       SD NAs
##  Overall 0.00  9.760  19.400  29.19 30.52581 39.8325 57.125 83.54 14.47232   0
##        0 0.00  9.502  19.165  28.78 29.82450 38.9050 54.515 76.56 13.87512   0
##        1 3.25 11.185  20.045  30.72 33.38965 44.2400 62.690 83.54 16.40269   0
## 
## Bootstrap summary: 
##           Variable  Min.    5% 1st Qu. Median  Mean 3rd Qu.   95%  Max.   SD
##   optimal_cutpoint 15.59 34.53   40.70  47.58 45.21   48.09 53.04 58.31 5.58
##              AUC_b  0.50  0.52    0.54   0.55  0.55    0.56  0.58  0.61 0.02
##            AUC_oob  0.48  0.52    0.54   0.55  0.55    0.57  0.59  0.63 0.02
##    sum_sens_spec_b  1.05  1.09    1.11   1.12  1.12    1.14  1.16  1.21 0.02
##  sum_sens_spec_oob  0.95  1.02    1.06   1.09  1.09    1.11  1.14  1.18 0.04
##              acc_b  0.34  0.62    0.70   0.76  0.74    0.77  0.79  0.82 0.06
##            acc_oob  0.26  0.59    0.69   0.75  0.72    0.77  0.79  0.82 0.06
##      sensitivity_b  0.09  0.16    0.21   0.24  0.27    0.33  0.48  0.92 0.10
##    sensitivity_oob  0.06  0.14    0.19   0.22  0.24    0.29  0.41  0.81 0.08
##      specificity_b  0.19  0.65    0.79   0.89  0.85    0.90  0.94  0.98 0.09
##    specificity_oob  0.13  0.63    0.78   0.89  0.84    0.90  0.94  0.98 0.10
##     cohens_kappa_b  0.03  0.08    0.11   0.13  0.13    0.15  0.17  0.22 0.03
##   cohens_kappa_oob -0.03  0.02    0.07   0.10  0.09    0.13  0.16  0.21 0.04
##              ppv_b  0.20  0.25    0.28   0.33  0.33    0.36  0.42  0.55 0.06
##            ppv_oob  0.15  0.20    0.25   0.30  0.30    0.34  0.40  0.56 0.06
##              npv_b  0.79  0.81    0.82   0.83  0.83    0.83  0.84  0.91 0.01
##            npv_oob  0.75  0.80    0.81   0.82  0.82    0.83  0.84  0.86 0.01
##  NAs
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0

# ROC plot
plot(PER.MAC.delaval.SCM)

# Youden index plot
plot_metric(PER.MAC.delaval.SCM)

# Summary of results when using SCM as reference

print(c("Optimal n NEU:", exp(NEU.delaval.SCM$optimal_cutpoint), NEU.delaval.SCM$sensitivity, NEU.delaval.SCM$specificity, NEU.delaval.SCM$ppv, NEU.delaval.SCM$npv, NEU.delaval.SCM$AUC))

##                                                 sensitivity         specificity 
##    "Optimal n NEU:"                "75" "0.924528301886792" "0.914191419141914" 
##                                                             
## "0.725158562367865" "0.980184005661713" "0.967137252808839"

print(c("Optimal %NEU:", PER.NEU.delaval.SCM$optimal_cutpoint, PER.NEU.delaval.SCM$sensitivity, PER.NEU.delaval.SCM$specificity, PER.NEU.delaval.SCM$ppv, PER.NEU.delaval.SCM$npv, PER.NEU.delaval.SCM$AUC))

##                                                 sensitivity         specificity 
##     "Optimal %NEU:"              "58.7" "0.347708894878706" "0.763036303630363" 
##                                                             
## "0.264344262295082"  "0.82689556509299" "0.514406696734363"

print(c("Optimal n LYM:", exp(LYM.delaval.SCM$optimal_cutpoint), LYM.delaval.SCM$sensitivity, LYM.delaval.SCM$specificity, LYM.delaval.SCM$ppv, LYM.delaval.SCM$npv, LYM.delaval.SCM$AUC))

##                                                 sensitivity         specificity 
##    "Optimal n LYM:"                "28" "0.886792452830189" "0.904950495049505" 
##                                                             
## "0.695560253699789" "0.970276008492569" "0.955351249410655"

print(c("Optimal %LYM:", PER.LYM.delaval.SCM$optimal_cutpoint, PER.LYM.delaval.SCM$sensitivity, PER.LYM.delaval.SCM$specificity, PER.LYM.delaval.SCM$ppv, PER.LYM.delaval.SCM$npv, PER.LYM.delaval.SCM$AUC))

##                                                 sensitivity         specificity 
##     "Optimal %LYM:"             "19.72" "0.450825082508251" "0.787061994609164" 
##                                                             
## "0.896325459317585" "0.259786476868327" "0.639895741595723"

print(c("Optimal n MAC:", exp(MAC.delaval.SCM$optimal_cutpoint), MAC.delaval.SCM$sensitivity, MAC.delaval.SCM$specificity, MAC.delaval.SCM$ppv, MAC.delaval.SCM$npv, MAC.delaval.SCM$AUC))

##                                                 sensitivity         specificity 
##    "Optimal n MAC:"                "54" "0.873315363881402" "0.903630363036304" 
##                                                             
##  "0.68936170212766"  "0.96680790960452" "0.946867355199132"

print(c("Optimal %MAC:", PER.MAC.delaval.SCM$optimal_cutpoint, PER.MAC.delaval.SCM$sensitivity, PER.MAC.delaval.SCM$specificity, PER.MAC.delaval.SCM$ppv, PER.MAC.delaval.SCM$npv, PER.MAC.delaval.SCM$AUC))

##                                                 sensitivity         specificity 
##     "Optimal %MAC:"             "47.98" "0.215633423180593" "0.897689768976898" 
##                                                             
## "0.340425531914894"  "0.82374318594791"  "0.55299564996931"

# CI using bootstrapping

# NEU.delaval.SCM
NEU.delaval.SCM_optimal_cutpoint <- boot_ci(NEU.delaval.SCM, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
NEU.delaval.SCM_lower_ci <- subset(NEU.delaval.SCM_optimal_cutpoint, NEU.delaval.SCM_optimal_cutpoint$quantile == 0.025)$values
exp(NEU.delaval.SCM_lower_ci)

## [1] 68

NEU.delaval.SCM_upper_ci <- subset(NEU.delaval.SCM_optimal_cutpoint, NEU.delaval.SCM_optimal_cutpoint$quantile == 0.975)$values
exp(NEU.delaval.SCM_upper_ci)

## [1] 80

boot_ci(NEU.delaval.SCM, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.895
## 2    0.975  0.952

boot_ci(NEU.delaval.SCM, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.892
## 2    0.975  0.935

boot_ci(NEU.delaval.SCM, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.673
## 2    0.975  0.781

boot_ci(NEU.delaval.SCM, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.972
## 2    0.975  0.987

boot_ci(NEU.delaval.SCM, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.956
## 2    0.975  0.977

# PER.NEU.delaval.SCM
PER.NEU.delaval.SCM_optimal_cutpoint <- boot_ci(PER.NEU.delaval.SCM, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
PER.NEU.delaval.SCM_optimal_cutpoint

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025   57.3
## 2    0.975   65.2

PER.NEU.delaval.SCM_lower_ci <- subset(PER.NEU.delaval.SCM_optimal_cutpoint, PER.NEU.delaval.SCM_optimal_cutpoint$quantile == 0.025)$values
PER.NEU.delaval.SCM_upper_ci <- subset(PER.NEU.delaval.SCM_optimal_cutpoint, PER.NEU.delaval.SCM_optimal_cutpoint$quantile == 0.975)$values

boot_ci(PER.NEU.delaval.SCM, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.195
## 2    0.975  0.410

boot_ci(PER.NEU.delaval.SCM, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.730
## 2    0.975  0.903

boot_ci(PER.NEU.delaval.SCM, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.238
## 2    0.975  0.357

boot_ci(PER.NEU.delaval.SCM, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.808
## 2    0.975  0.847

boot_ci(PER.NEU.delaval.SCM, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.479
## 2    0.975  0.551

# LYM.delaval.SCM
LYM.delaval.SCM_optimal_cutpoint <- boot_ci(LYM.delaval.SCM, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
LYM.delaval.SCM_lower_ci <- subset(LYM.delaval.SCM_optimal_cutpoint, LYM.delaval.SCM_optimal_cutpoint$quantile == 0.025)$values
exp(LYM.delaval.SCM_lower_ci)

## [1] 25

LYM.delaval.SCM_upper_ci <- subset(LYM.delaval.SCM_optimal_cutpoint, LYM.delaval.SCM_optimal_cutpoint$quantile == 0.975)$values
exp(LYM.delaval.SCM_upper_ci)

## [1] 33

boot_ci(LYM.delaval.SCM, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.851
## 2    0.975  0.932

boot_ci(LYM.delaval.SCM, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.865
## 2    0.975  0.939

boot_ci(LYM.delaval.SCM, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.617
## 2    0.975  0.773

boot_ci(LYM.delaval.SCM, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.962
## 2    0.975  0.982

boot_ci(LYM.delaval.SCM, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.942
## 2    0.975  0.967

# PER.LYM.delaval.SCM
PER.LYM.delaval.SCM_optimal_cutpoint <- boot_ci(PER.LYM.delaval.SCM, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
PER.LYM.delaval.SCM_optimal_cutpoint

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025   18.4
## 2    0.975   21.8

PER.LYM.delaval.SCM_lower_ci <- subset(PER.LYM.delaval.SCM_optimal_cutpoint, PER.LYM.delaval.SCM_optimal_cutpoint$quantile == 0.025)$values
PER.LYM.delaval.SCM_upper_ci <- subset(PER.LYM.delaval.SCM_optimal_cutpoint, PER.LYM.delaval.SCM_optimal_cutpoint$quantile == 0.975)$values

boot_ci(PER.LYM.delaval.SCM, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.356
## 2    0.975  0.513

boot_ci(PER.LYM.delaval.SCM, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.725
## 2    0.975  0.884

boot_ci(PER.LYM.delaval.SCM, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.877
## 2    0.975  0.929

boot_ci(PER.LYM.delaval.SCM, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.234
## 2    0.975  0.285

boot_ci(PER.LYM.delaval.SCM, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.612
## 2    0.975  0.669

# MAC.delaval.SCM
MAC.delaval.SCM_optimal_cutpoint <- boot_ci(MAC.delaval.SCM, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
MAC.delaval.SCM_lower_ci <- subset(MAC.delaval.SCM_optimal_cutpoint, MAC.delaval.SCM_optimal_cutpoint$quantile == 0.025)$values
exp(LYM.delaval.SCM_upper_ci)

## [1] 33

MAC.delaval.SCM_upper_ci <- subset(MAC.delaval.SCM_optimal_cutpoint, MAC.delaval.SCM_optimal_cutpoint$quantile == 0.975)$values
exp(MAC.delaval.SCM_upper_ci)

## [1] 60

boot_ci(MAC.delaval.SCM, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.846
## 2    0.975  0.924

boot_ci(MAC.delaval.SCM, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.865
## 2    0.975  0.921

boot_ci(MAC.delaval.SCM, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.614
## 2    0.975  0.737

boot_ci(MAC.delaval.SCM, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.960
## 2    0.975  0.980

boot_ci(MAC.delaval.SCM, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.933
## 2    0.975  0.960

# PER.MAC.delaval.SCM
PER.MAC.delaval.SCM_optimal_cutpoint <- boot_ci(PER.MAC.delaval.SCM, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
PER.MAC.delaval.SCM_optimal_cutpoint

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025   33.5
## 2    0.975   53.3

PER.MAC.delaval.SCM_lower_ci <- subset(PER.MAC.delaval.SCM_optimal_cutpoint, PER.MAC.delaval.SCM_optimal_cutpoint$quantile == 0.025)$values
PER.MAC.delaval.SCM_upper_ci <- subset(PER.MAC.delaval.SCM_optimal_cutpoint, PER.MAC.delaval.SCM_optimal_cutpoint$quantile == 0.975)$values

boot_ci(PER.MAC.delaval.SCM, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.145
## 2    0.975  0.500

boot_ci(PER.MAC.delaval.SCM, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.627
## 2    0.975  0.950

boot_ci(PER.MAC.delaval.SCM, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.237
## 2    0.975  0.450

boot_ci(PER.MAC.delaval.SCM, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.808
## 2    0.975  0.847

boot_ci(PER.MAC.delaval.SCM, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.518
## 2    0.975  0.586

# Prevalence

# True prevalence

table(Doff.by.quarter.2$delaval.SCM)

## 
##    0    1 
## 1529  372

prop.table(table(Doff.by.quarter.2$delaval.SCM))

## 
##         0         1 
## 0.8043135 0.1956865

# ---- Neutrophils ----

# optimal cut-off

optimal_neu_doff_delaval.SCM_quarter <- NEU.delaval.SCM$optimal_cutpoint

table(Doff.by.quarter.2$TOTAL01_N > exp(optimal_neu_doff_delaval.SCM_quarter))

## 
## FALSE  TRUE 
##  1417   473

prop.table(table(Doff.by.quarter.2$TOTAL01_N > exp(optimal_neu_doff_delaval.SCM_quarter)))

## 
##     FALSE      TRUE 
## 0.7497354 0.2502646

# confidence intervals

table(Doff.by.quarter.2$TOTAL01_N > exp(NEU.delaval.SCM_lower_ci))

## 
## FALSE  TRUE 
##  1381   509

prop.table(table(Doff.by.quarter.2$TOTAL01_N > exp(NEU.delaval.SCM_lower_ci)))

## 
##     FALSE      TRUE 
## 0.7306878 0.2693122

table(Doff.by.quarter.2$TOTAL01_N > exp(NEU.delaval.SCM_upper_ci))

## 
## FALSE  TRUE 
##  1440   450

prop.table(table(Doff.by.quarter.2$TOTAL01_N >exp(NEU.delaval.SCM_upper_ci)))

## 
##     FALSE      TRUE 
## 0.7619048 0.2380952

# ---- Percent neu ----

# Optimal cut-off

optimal_per_neu_doff_delaval.SCM_quarter <- PER.NEU.delaval.SCM$optimal_cutpoint

table(Doff.by.quarter.2$PERCENT01_N > optimal_per_neu_doff_delaval.SCM_quarter)

## 
## FALSE  TRUE 
##  1405   485

prop.table(table(Doff.by.quarter.2$PERCENT01_N > optimal_per_neu_doff_delaval.SCM_quarter))

## 
##     FALSE      TRUE 
## 0.7433862 0.2566138

# confidence intervals

table(Doff.by.quarter.2$PERCENT01_N > PER.NEU.delaval.SCM_lower_ci)

## 
## FALSE  TRUE 
##  1342   548

prop.table(table(Doff.by.quarter.2$PERCENT01_N > PER.NEU.delaval.SCM_lower_ci))

## 
##     FALSE      TRUE 
## 0.7100529 0.2899471

table(Doff.by.quarter.2$PERCENT01_N > PER.NEU.delaval.SCM_upper_ci)

## 
## FALSE  TRUE 
##  1675   215

prop.table(table(Doff.by.quarter.2$PERCENT01_N >PER.NEU.delaval.SCM_upper_ci))

## 
##     FALSE      TRUE 
## 0.8862434 0.1137566

# ---- Lymphocytes ----

# optimal cut-off

optimal_lym_doff_delaval.SCM_quarter <-LYM.delaval.SCM$optimal_cutpoint

table(Doff.by.quarter.2$TOTAL02_N > exp(optimal_lym_doff_delaval.SCM_quarter))

## 
## FALSE  TRUE 
##  1417   473

prop.table(table(Doff.by.quarter.2$TOTAL02_N > exp(optimal_lym_doff_delaval.SCM_quarter)))

## 
##     FALSE      TRUE 
## 0.7497354 0.2502646

# confidence intervals

table(Doff.by.quarter.2$TOTAL02_N > exp(LYM.delaval.SCM_lower_ci))

## 
## FALSE  TRUE 
##  1362   528

prop.table(table(Doff.by.quarter.2$TOTAL02_N > exp(LYM.delaval.SCM_lower_ci)))

## 
##     FALSE      TRUE 
## 0.7206349 0.2793651

table(Doff.by.quarter.2$TOTAL02_N > exp(LYM.delaval.SCM_upper_ci))

## 
## FALSE  TRUE 
##  1504   386

prop.table(table(Doff.by.quarter.2$TOTAL02_N >exp(LYM.delaval.SCM_upper_ci)))

## 
##     FALSE      TRUE 
## 0.7957672 0.2042328

# ---- Percent lym ----

# optimal cut-off

optimal_per_lym_doff_delaval.SCM_quarter <-PER.LYM.delaval.SCM$optimal_cutpoint

table(Doff.by.quarter.2$PERCENT02_N > optimal_per_lym_doff_delaval.SCM_quarter)

## 
## FALSE  TRUE 
##  1129   761

prop.table(table(Doff.by.quarter.2$PERCENT02_N > optimal_per_lym_doff_delaval.SCM_quarter))

## 
##     FALSE      TRUE 
## 0.5973545 0.4026455

# confidence intervals

table(Doff.by.quarter.2$PERCENT02_N > PER.LYM.delaval.SCM_lower_ci)

## 
## FALSE  TRUE 
##   997   893

prop.table(table(Doff.by.quarter.2$PERCENT02_N > PER.LYM.delaval.SCM_lower_ci))

## 
##     FALSE      TRUE 
## 0.5275132 0.4724868

table(Doff.by.quarter.2$PERCENT02_N > PER.LYM.delaval.SCM_upper_ci)

## 
## FALSE  TRUE 
##  1293   597

prop.table(table(Doff.by.quarter.2$PERCENT02_N >PER.LYM.delaval.SCM_upper_ci))

## 
##    FALSE     TRUE 
## 0.684127 0.315873

# ---- Macrophages ----

# optimal cut-off

optimal_mac_doff_delaval.SCM_quarter <- MAC.delaval.SCM$optimal_cutpoint

table(Doff.by.quarter.2$TOTAL03_N > exp(optimal_mac_doff_delaval.SCM_quarter))

## 
## FALSE  TRUE 
##  1428   462

prop.table(table(Doff.by.quarter.2$TOTAL03_N > exp(optimal_mac_doff_delaval.SCM_quarter)))

## 
##     FALSE      TRUE 
## 0.7555556 0.2444444

# confidence intervals

table(Doff.by.quarter.2$TOTAL03_N > exp(MAC.delaval.SCM_lower_ci))

## 
## FALSE  TRUE 
##  1381   509

prop.table(table(Doff.by.quarter.2$TOTAL03_N > exp(MAC.delaval.SCM_lower_ci)))

## 
##     FALSE      TRUE 
## 0.7306878 0.2693122

table(Doff.by.quarter.2$TOTAL03_N > exp(MAC.delaval.SCM_upper_ci))

## 
## FALSE  TRUE 
##  1455   435

prop.table(table(Doff.by.quarter.2$TOTAL03_N >exp(MAC.delaval.SCM_upper_ci)))

## 
##     FALSE      TRUE 
## 0.7698413 0.2301587

# ---- Percent MAC -----

# optimal cut-off

optimal_per_mac_doff_delaval.SCM_quarter <- PER.MAC.delaval.SCM$optimal_cutpoint

table(Doff.by.quarter.2$PERCENT03_N > optimal_per_mac_doff_delaval.SCM_quarter)

## 
## FALSE  TRUE 
##  1655   235

prop.table(table(Doff.by.quarter.2$PERCENT03_N > optimal_per_mac_doff_delaval.SCM_quarter))

## 
##     FALSE      TRUE 
## 0.8756614 0.1243386

# confidence intervals

table(Doff.by.quarter.2$PERCENT03_N > PER.MAC.delaval.SCM_lower_ci)

## 
## FALSE  TRUE 
##  1133   757

prop.table(table(Doff.by.quarter.2$PERCENT03_N > PER.MAC.delaval.SCM_lower_ci))

## 
##     FALSE      TRUE 
## 0.5994709 0.4005291

table(Doff.by.quarter.2$PERCENT03_N > PER.MAC.delaval.SCM_upper_ci)

## 
## FALSE  TRUE 
##  1751   139

prop.table(table(Doff.by.quarter.2$PERCENT03_N >PER.MAC.delaval.SCM_upper_ci))

## 
##      FALSE       TRUE 
## 0.92645503 0.07354497

Postpartum: C3

Cow IMI

# Set a seed for reproducibility

set.seed(123)

# IMI cow referent

# SCC IMI cow
SCC.IMI <- cutpointr(Pc.by.cow.C3, log.SCC.CO.COW, IMI.cow, 
                           method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 0

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(SCC.IMI)

## Method: maximize_metric 
## Predictor: log.SCC.CO.COW 
## Outcome: IMI.cow 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC   n n_pos n_neg
##  0.4822 425   114   311
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv tp
##            4.6151        1.0318 0.4541      0.6491      0.3826 0.2782 0.7484 74
##  fn  fp  tn
##  40 192 119
## 
## Predictor summary: 
##     Data     Min.       5%  1st Qu.   Median     Mean  3rd Qu.      95%
##  Overall 2.833213 3.610918 4.356709 4.875197 5.006120 5.420535 7.305560
##        0 3.258097 3.465736 4.314000 4.908910 4.905444 5.316837 6.212545
##        1 2.833213 3.676221 4.375737 4.875197 5.043024 5.457180 7.536019
##      Max.        SD NAs
##  8.253228 1.0056560   0
##  7.658228 0.8688883   0
##  8.253228 1.0502220   0
## 
## Bootstrap summary: 
##           Variable  Min.    5% 1st Qu. Median  Mean 3rd Qu.  95% Max.   SD NAs
##   optimal_cutpoint  3.26  3.26    4.50   4.62   Inf    4.97 5.43  Inf  NaN   0
##              AUC_b  0.37  0.43    0.46   0.48  0.48    0.50 0.53 0.59 0.03   0
##            AUC_oob  0.35  0.41    0.45   0.48  0.48    0.51 0.55 0.63 0.04   0
##    sum_sens_spec_b  0.98  1.01    1.03   1.06  1.06    1.09 1.14 1.22 0.04   0
##  sum_sens_spec_oob  0.70  0.86    0.92   0.97  0.96    1.01 1.05 1.14 0.06   0
##              acc_b  0.21  0.27    0.43   0.48  0.47    0.55 0.63 0.77 0.11   0
##            acc_oob  0.16  0.26    0.37   0.45  0.43    0.50 0.57 0.77 0.10   0
##      sensitivity_b  0.00  0.33    0.52   0.67  0.67    0.76 1.00 1.00 0.21   0
##    sensitivity_oob  0.00  0.21    0.42   0.57  0.59    0.71 1.00 1.00 0.24   0
##      specificity_b  0.00  0.01    0.32   0.41  0.40    0.56 0.74 1.00 0.23   0
##    specificity_oob  0.00  0.01    0.28   0.39  0.37    0.53 0.70 1.00 0.22   0
##     cohens_kappa_b -0.02  0.01    0.02   0.04  0.05    0.07 0.11 0.18 0.03   0
##   cohens_kappa_oob -0.20 -0.11   -0.06  -0.02 -0.03    0.00 0.04 0.11 0.05   0
##              ppv_b  0.20  0.25    0.27   0.29  0.29    0.31 0.35 0.42 0.03   2
##            ppv_oob  0.00  0.18    0.23   0.25  0.25    0.28 0.31 0.38 0.04   2
##              npv_b  0.68  0.72    0.75   0.77  0.81    0.81 1.00 1.00 0.09   0
##            npv_oob  0.00  0.60    0.68   0.72  0.73    0.75 1.00 1.00 0.13  29

# ROC plot
plot(SCC.IMI)

## Warning: Removed 2 rows containing non-finite outside the scale range
## (`stat_density()`).

# Youden index plot
plot_metric(SCC.IMI)

# NEU IMI cow
NEU.IMI <- cutpointr(Pc.by.cow.C3, log.TOTAL.NEU.CO.COW, IMI.cow, 
                           method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(NEU.IMI)

## Method: maximize_metric 
## Predictor: log.TOTAL.NEU.CO.COW 
## Outcome: IMI.cow 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC   n n_pos n_neg
##  0.5446 427   312   115
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv
##            4.5326         1.083 0.4543      0.3526      0.7304 0.7801 0.2937
##   tp  fn fp tn
##  110 202 31 84
## 
## Predictor summary: 
##     Data     Min.       5%  1st Qu.   Median     Mean  3rd Qu.      95%
##  Overall 2.197225 2.890372 3.569433 4.127134 4.266303 4.740565 6.127288
##        0 2.564949 2.855037 3.496048 4.077537 4.121409 4.614924 5.634352
##        1 2.197225 2.920109 3.583519 4.166635 4.319710 4.781216 6.235343
##      Max.        SD NAs
##  8.663369 1.0412536   0
##  7.411556 0.9043542   0
##  8.663369 1.0838380   0
## 
## Bootstrap summary: 
##           Variable  Min.    5% 1st Qu. Median Mean 3rd Qu.  95% Max.   SD NAs
##   optimal_cutpoint  2.77  3.33    3.96   4.53 4.39    4.84 5.47 7.85 0.70   0
##              AUC_b  0.43  0.49    0.52   0.54 0.54    0.57 0.59 0.68 0.03   0
##            AUC_oob  0.40  0.48    0.52   0.54 0.55    0.57 0.61 0.69 0.04   0
##    sum_sens_spec_b  1.02  1.06    1.09   1.12 1.12    1.14 1.19 1.32 0.04   0
##  sum_sens_spec_oob  0.77  0.91    0.97   1.01 1.01    1.05 1.10 1.23 0.06   0
##              acc_b  0.25  0.35    0.41   0.49 0.51    0.60 0.71 0.77 0.12   0
##            acc_oob  0.20  0.31    0.37   0.45 0.47    0.55 0.68 0.75 0.12   0
##      sensitivity_b  0.02  0.13    0.25   0.39 0.46    0.63 0.88 0.99 0.26   0
##    sensitivity_oob  0.00  0.11    0.21   0.36 0.43    0.60 0.87 0.98 0.26   0
##      specificity_b  0.06  0.22    0.50   0.74 0.66    0.88 0.96 1.00 0.25   0
##    specificity_oob  0.00  0.13    0.36   0.64 0.58    0.80 0.94 1.00 0.27   0
##     cohens_kappa_b  0.01  0.04    0.06   0.09 0.09    0.12 0.16 0.25 0.04   0
##   cohens_kappa_oob -0.18 -0.08   -0.02   0.01 0.01    0.03 0.07 0.18 0.05   0
##              ppv_b  0.69  0.74    0.77   0.80 0.81    0.85 0.91 1.00 0.05   0
##            ppv_oob  0.52  0.67    0.71   0.74 0.75    0.78 0.87 1.00 0.06   2
##              npv_b  0.22  0.27    0.29   0.32 0.33    0.35 0.44 0.82 0.05   0
##            npv_oob  0.00  0.20    0.25   0.27 0.27    0.30 0.34 0.50 0.04   0

# ROC plot
plot(NEU.IMI)

# Youden index plot
plot_metric(NEU.IMI)

# %NEU IMI cow
PER.NEU.IMI <- cutpointr(Pc.by.cow.C3, PERCENT.NEU.CO.COW, IMI.cow, 
                               method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(PER.NEU.IMI)

## Method: maximize_metric 
## Predictor: PERCENT.NEU.CO.COW 
## Outcome: IMI.cow 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC   n n_pos n_neg
##  0.5584 427   312   115
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv
##              60.4        1.1229 0.5035      0.4359       0.687 0.7907 0.3098
##   tp  fn fp tn
##  136 176 36 79
## 
## Predictor summary: 
##     Data  Min.     5% 1st Qu. Median     Mean 3rd Qu.    95%  Max.       SD NAs
##  Overall 21.23 36.928 48.7900  57.41 56.97274  65.370 75.682 86.98 11.78689   0
##        0 24.42 36.453 46.7650  55.42 55.29061  62.955 75.129 81.86 12.02568   0
##        1 21.23 37.723 49.6975  58.11 57.59276  65.850 75.805 86.98 11.65595   0
## 
## Bootstrap summary: 
##           Variable  Min.    5% 1st Qu. Median  Mean 3rd Qu.   95%  Max.   SD
##   optimal_cutpoint 38.36 45.87   54.19  57.60 56.89   60.69 64.38 77.78 5.83
##              AUC_b  0.45  0.51    0.54   0.56  0.56    0.58  0.61  0.67 0.03
##            AUC_oob  0.39  0.49    0.53   0.56  0.56    0.59  0.62  0.71 0.04
##    sum_sens_spec_b  0.91  1.08    1.12   1.15  1.15    1.18  1.23  1.29 0.05
##  sum_sens_spec_oob  0.79  0.95    1.01   1.05  1.05    1.10  1.16  1.26 0.06
##              acc_b  0.27  0.45    0.50   0.56  0.57    0.63  0.70  0.77 0.08
##            acc_oob  0.26  0.41    0.47   0.52  0.53    0.58  0.67  0.74 0.08
##      sensitivity_b  0.03  0.33    0.42   0.53  0.55    0.68  0.84  0.96 0.17
##    sensitivity_oob  0.01  0.29    0.39   0.49  0.53    0.65  0.82  0.97 0.17
##      specificity_b  0.08  0.30    0.47   0.64  0.60    0.74  0.82  1.00 0.17
##    specificity_oob  0.03  0.21    0.38   0.56  0.53    0.69  0.78  1.00 0.19
##     cohens_kappa_b -0.05  0.06    0.09   0.12  0.12    0.15  0.19  0.28 0.04
##   cohens_kappa_oob -0.17 -0.04    0.01   0.04  0.04    0.08  0.13  0.22 0.05
##              ppv_b  0.59  0.74    0.77   0.79  0.79    0.82  0.85  1.00 0.03
##            ppv_oob  0.50  0.69    0.73   0.75  0.76    0.78  0.82  1.00 0.04
##              npv_b  0.23  0.28    0.31   0.34  0.34    0.37  0.43  0.60 0.05
##            npv_oob  0.10  0.22    0.26   0.29  0.29    0.32  0.37  0.50 0.05
##  NAs
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0

# ROC plot
plot(PER.NEU.IMI)

# Youden index plot
plot_metric(PER.NEU.IMI)

# LYM IMI cow

LYM.IMI <-  cutpointr(Pc.by.cow.C3, log.TOTAL.LYM.CO.COW, IMI.cow, 
                            method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 0

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(LYM.IMI)

## Method: maximize_metric 
## Predictor: log.TOTAL.LYM.CO.COW 
## Outcome: IMI.cow 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC   n n_pos n_neg
##  0.4718 427   115   312
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv tp
##             3.091        1.0281 0.5105      0.5217      0.5064 0.2804 0.7418 60
##  fn  fp  tn
##  55 154 158
## 
## Predictor summary: 
##     Data     Min.       5%  1st Qu.   Median     Mean  3rd Qu.      95%
##  Overall 0.000000 1.791759 2.484907 3.091042 3.161188 3.610918 4.953707
##        0 0.000000 1.609438 2.397895 3.091042 3.041403 3.526361 4.409734
##        1 1.098612 1.945910 2.564949 3.044522 3.205340 3.637586 4.993262
##      Max.        SD NAs
##  6.921658 0.9859764   0
##  6.556778 0.9596418   0
##  6.921658 0.9933758   0
## 
## Bootstrap summary: 
##           Variable  Min.    5% 1st Qu. Median  Mean 3rd Qu.  95% Max.   SD NAs
##   optimal_cutpoint  1.61  2.71    3.09   3.26   Inf    3.50  Inf  Inf  NaN   0
##              AUC_b  0.36  0.42    0.45   0.47  0.47    0.49 0.52 0.58 0.03   0
##            AUC_oob  0.33  0.41    0.44   0.47  0.47    0.50 0.54 0.66 0.04   0
##    sum_sens_spec_b  0.98  1.00    1.02   1.05  1.05    1.08 1.13 1.19 0.04   0
##  sum_sens_spec_oob  0.73  0.85    0.93   0.98  0.97    1.00 1.06 1.18 0.06   0
##              acc_b  0.22  0.42    0.53   0.56  0.58    0.62 0.74 0.79 0.10   0
##            acc_oob  0.22  0.36    0.49   0.52  0.54    0.58 0.74 0.81 0.11   0
##      sensitivity_b  0.00  0.00    0.35   0.48  0.42    0.55 0.74 1.00 0.24   0
##    sensitivity_oob  0.00  0.00    0.22   0.41  0.35    0.49 0.64 0.98 0.22   0
##      specificity_b  0.01  0.32    0.52   0.59  0.64    0.71 1.00 1.00 0.22   0
##    specificity_oob  0.00  0.28    0.50   0.56  0.61    0.68 1.00 1.00 0.23   0
##     cohens_kappa_b -0.02  0.00    0.02   0.04  0.04    0.07 0.11 0.16 0.03   0
##   cohens_kappa_oob -0.22 -0.12   -0.06  -0.02 -0.03    0.00 0.05 0.13 0.05   0
##              ppv_b  0.22  0.25    0.28   0.30  0.34    0.33 0.67 1.00 0.16 132
##            ppv_oob  0.00  0.00    0.22   0.25  0.23    0.28 0.32 0.39 0.08 249
##              npv_b  0.67  0.71    0.73   0.75  0.75    0.77 0.80 1.00 0.04   0
##            npv_oob  0.00  0.62    0.69   0.72  0.71    0.74 0.78 0.85 0.08   0

# ROC plot
plot(LYM.IMI)

## Warning: Removed 132 rows containing non-finite outside the scale range
## (`stat_density()`).

# Youden index plot
plot_metric(LYM.IMI)

# %LYM IMI cow
PER.LYM.IMI <- cutpointr(Pc.by.cow.C3, PERCENT.LYM.CO.COW, IMI.cow, 
                               method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 0

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(PER.LYM.IMI)

## Method: maximize_metric 
## Predictor: PERCENT.LYM.CO.COW 
## Outcome: IMI.cow 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC   n n_pos n_neg
##  0.4985 427   115   312
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv
##             13.13        1.0559 0.3583      0.8957      0.1603 0.2822 0.8065
##   tp fn  fp tn
##  103 12 262 50
## 
## Predictor summary: 
##     Data Min.      5% 1st Qu. Median     Mean 3rd Qu.    95%  Max.       SD NAs
##  Overall 5.08 10.1460  15.675  19.20 19.32700 22.7400 29.383 51.10 5.899042   0
##        0 5.71  9.7530  16.155  19.30 19.23061 22.3750 28.473 35.94 5.629593   0
##        1 5.08 10.5395  15.520  19.19 19.36253 22.8475 29.465 51.10 6.003713   0
## 
## Bootstrap summary: 
##           Variable  Min.    5% 1st Qu. Median  Mean 3rd Qu.   95%  Max.   SD
##   optimal_cutpoint  5.71 12.38   13.68  16.14 17.23   18.29 27.00 35.94 4.48
##              AUC_b  0.40  0.45    0.48   0.50  0.50    0.52  0.55  0.59 0.03
##            AUC_oob  0.34  0.43    0.47   0.50  0.50    0.53  0.57  0.66 0.04
##    sum_sens_spec_b  0.93  1.02    1.06   1.08  1.08    1.11  1.15  1.23 0.04
##  sum_sens_spec_oob  0.78  0.89    0.95   1.00  0.99    1.04  1.09  1.19 0.06
##              acc_b  0.24  0.34    0.38   0.45  0.47    0.52  0.71  0.79 0.11
##            acc_oob  0.18  0.31    0.36   0.41  0.44    0.48  0.68  0.76 0.11
##      sensitivity_b  0.02  0.13    0.61   0.76  0.69    0.88  0.93  1.00 0.24
##    sensitivity_oob  0.00  0.07    0.51   0.69  0.62    0.84  0.92  1.00 0.26
##      specificity_b  0.00  0.14    0.19   0.32  0.39    0.48  0.92  0.99 0.24
##    specificity_oob  0.00  0.12    0.17   0.30  0.37    0.45  0.91  0.98 0.24
##     cohens_kappa_b -0.05  0.02    0.04   0.06  0.06    0.08  0.11  0.19 0.03
##   cohens_kappa_oob -0.20 -0.10   -0.04   0.00 -0.01    0.02  0.06  0.12 0.05
##              ppv_b  0.21  0.25    0.28   0.30  0.31    0.32  0.39  0.67 0.05
##            ppv_oob  0.00  0.18    0.24   0.26  0.26    0.29  0.32  0.50 0.05
##              npv_b  0.66  0.73    0.76   0.79  0.80    0.82  0.88  1.00 0.05
##            npv_oob  0.00  0.65    0.70   0.73  0.73    0.77  0.85  1.00 0.07
##  NAs
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    5

# ROC plot
plot(PER.LYM.IMI)

# Youden index plot
plot_metric(PER.LYM.IMI)

# MAC IMI cow
MAC.IMI <- cutpointr(Pc.by.cow.C3, log.TOTAL.MAC.CO.COW, IMI.cow, 
                           method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 0

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(MAC.IMI)

## Method: maximize_metric 
## Predictor: log.TOTAL.MAC.CO.COW 
## Outcome: IMI.cow 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC   n n_pos n_neg
##  0.4982 427   115   312
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv tp
##            3.0445        1.0313 0.4848      0.5826      0.4487 0.2803 0.7447 67
##  fn  fp  tn
##  48 172 140
## 
## Predictor summary: 
##     Data     Min.       5%  1st Qu.   Median     Mean  3rd Qu.      95%
##  Overall 0.000000 1.791759 2.564949 3.178054 3.286585 3.871201 5.385332
##        0 1.098612 1.899665 2.602003 3.178054 3.244394 3.860674 4.598808
##        1 0.000000 1.609438 2.564949 3.135494 3.302136 3.871201 5.459102
##      Max.       SD NAs
##  8.318254 1.124299   0
##  7.097549 0.960287   0
##  8.318254 1.180058   0
## 
## Bootstrap summary: 
##           Variable  Min.    5% 1st Qu. Median  Mean 3rd Qu.  95% Max.   SD NAs
##   optimal_cutpoint  1.10  1.79    2.20   3.04   Inf    3.64 4.13  Inf  NaN   0
##              AUC_b  0.40  0.45    0.48   0.50  0.50    0.52 0.55 0.61 0.03   0
##            AUC_oob  0.37  0.43    0.47   0.50  0.50    0.53 0.57 0.63 0.04   0
##    sum_sens_spec_b  0.95  1.02    1.05   1.07  1.07    1.09 1.14 1.25 0.04   0
##  sum_sens_spec_oob  0.75  0.87    0.93   0.97  0.97    1.01 1.05 1.15 0.06   0
##              acc_b  0.25  0.29    0.36   0.49  0.48    0.60 0.68 0.78 0.13   0
##            acc_oob  0.21  0.27    0.33   0.44  0.44    0.54 0.65 0.78 0.12   0
##      sensitivity_b  0.00  0.22    0.41   0.65  0.65    0.91 0.99 1.00 0.27   0
##    sensitivity_oob  0.00  0.13    0.30   0.57  0.58    0.85 0.98 1.00 0.29   0
##      specificity_b  0.00  0.05    0.15   0.45  0.42    0.67 0.84 1.00 0.27   0
##    specificity_oob  0.00  0.03    0.13   0.40  0.39    0.64 0.82 1.00 0.27   0
##     cohens_kappa_b -0.04  0.01    0.03   0.05  0.06    0.08 0.12 0.19 0.03   0
##   cohens_kappa_oob -0.23 -0.11   -0.05  -0.02 -0.03    0.01 0.03 0.14 0.05   0
##              ppv_b  0.20  0.25    0.28   0.30  0.30    0.32 0.37 1.00 0.04   2
##            ppv_oob  0.00  0.17    0.23   0.26  0.25    0.28 0.31 0.38 0.05   3
##              npv_b  0.63  0.72    0.75   0.78  0.80    0.83 0.94 1.00 0.07   0
##            npv_oob  0.29  0.62    0.68   0.72  0.72    0.75 0.86 1.00 0.08   6

# ROC plot
plot(MAC.IMI)

## Warning: Removed 2 rows containing non-finite outside the scale range
## (`stat_density()`).

# Youden index plot
plot_metric(MAC.IMI)

# %MAC IMI cow
PER.MAC.IMI<- cutpointr(Pc.by.cow.C3, PERCENT.MAC.CO.COW, IMI.cow, 
                              method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 0

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(PER.MAC.IMI)

## Method: maximize_metric 
## Predictor: PERCENT.MAC.CO.COW 
## Outcome: IMI.cow 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC   n n_pos n_neg
##  0.5549 427   115   312
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv tp
##             36.61        1.1075 0.7049      0.2261      0.8814 0.4127 0.7555 26
##  fn fp  tn
##  89 37 275
## 
## Predictor summary: 
##     Data Min.    5% 1st Qu. Median     Mean 3rd Qu.     95%  Max.       SD NAs
##  Overall 3.51 8.640 15.1500  21.96 23.69944 30.7300 44.7900 73.11 11.55920   0
##        0 6.06 8.602 16.7250  23.42 25.47730 34.4400 46.0650 59.41 12.33271   0
##        1 3.51 8.640 14.9075  21.56 23.04413 29.9475 42.3765 73.11 11.21031   0
## 
## Bootstrap summary: 
##           Variable  Min.    5% 1st Qu. Median  Mean 3rd Qu.   95%  Max.   SD
##   optimal_cutpoint  7.72 14.95   24.10  27.78 28.57   36.61 38.24 59.12 7.23
##              AUC_b  0.44  0.50    0.53   0.56  0.56    0.58  0.61  0.65 0.03
##            AUC_oob  0.41  0.49    0.53   0.55  0.55    0.58  0.62  0.70 0.04
##    sum_sens_spec_b  1.03  1.08    1.11   1.14  1.14    1.17  1.21  1.29 0.04
##  sum_sens_spec_oob  0.79  0.94    1.00   1.04  1.04    1.08  1.14  1.28 0.06
##              acc_b  0.28  0.43    0.59   0.65  0.63    0.70  0.74  0.78 0.09
##            acc_oob  0.25  0.38    0.54   0.60  0.59    0.67  0.72  0.80 0.10
##      sensitivity_b  0.03  0.20    0.28   0.42  0.44    0.55  0.84  0.99 0.19
##    sensitivity_oob  0.00  0.12    0.22   0.33  0.36    0.47  0.75  0.98 0.18
##      specificity_b  0.05  0.28    0.61   0.72  0.70    0.86  0.91  1.00 0.19
##    specificity_oob  0.01  0.25    0.57   0.70  0.68    0.85  0.91  1.00 0.19
##     cohens_kappa_b  0.02  0.06    0.10   0.14  0.14    0.17  0.21  0.29 0.05
##   cohens_kappa_oob -0.16 -0.05    0.00   0.04  0.04    0.08  0.14  0.33 0.06
##              ppv_b  0.23  0.29    0.33   0.37  0.38    0.42  0.52  0.80 0.07
##            ppv_oob  0.00  0.21    0.26   0.30  0.31    0.34  0.44  0.65 0.07
##              npv_b  0.69  0.73    0.76   0.77  0.78    0.79  0.83  0.95 0.03
##            npv_oob  0.50  0.68    0.72   0.74  0.74    0.77  0.80  0.87 0.04
##  NAs
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    1
##    0
##    0

# ROC plot
plot(PER.MAC.IMI)

# Youden index plot
plot_metric(PER.MAC.IMI)

# Summary of results when using C3 and  cow-level IMI as reference

print(c("Optimal SCC:", exp(SCC.IMI$optimal_cutpoint), SCC.IMI$sensitivity, SCC.IMI$specificity, SCC.IMI$ppv, SCC.IMI$npv, SCC.IMI$AUC))

##                                                 sensitivity         specificity 
##      "Optimal SCC:"               "101" "0.649122807017544" "0.382636655948553" 
##                                                             
## "0.278195488721804" "0.748427672955975" "0.482230495853782"

print(c("Optimal n NEU:", exp(NEU.IMI$optimal_cutpoint), NEU.IMI$sensitivity, NEU.IMI$specificity, NEU.IMI$ppv, NEU.IMI$npv, NEU.IMI$AUC))

##                                                 sensitivity         specificity 
##    "Optimal n NEU:"                "93" "0.352564102564103" "0.730434782608696" 
##                                                             
## "0.780141843971631" "0.293706293706294" "0.544648829431438"

print(c("Optimal %NEU:", PER.NEU.IMI$optimal_cutpoint, PER.NEU.IMI$sensitivity, PER.NEU.IMI$specificity, PER.NEU.IMI$ppv, PER.NEU.IMI$npv, PER.NEU.IMI$AUC))

##                                                 sensitivity         specificity 
##     "Optimal %NEU:"              "60.4" "0.435897435897436"  "0.68695652173913" 
##                                                             
## "0.790697674418605" "0.309803921568627" "0.558389074693422"

print(c("Optimal n LYM:", exp(LYM.IMI$optimal_cutpoint), LYM.IMI$sensitivity, LYM.IMI$specificity, LYM.IMI$ppv, LYM.IMI$npv, LYM.IMI$AUC))

##                                                 sensitivity         specificity 
##    "Optimal n LYM:"                "22" "0.521739130434783" "0.506410256410256" 
##                                                             
## "0.280373831775701" "0.741784037558685" "0.471794871794872"

print(c("Optimal %LYM:", PER.LYM.IMI$optimal_cutpoint, PER.LYM.IMI$sensitivity, PER.LYM.IMI$specificity, PER.LYM.IMI$ppv, PER.LYM.IMI$npv, PER.LYM.IMI$AUC))

##                                                 sensitivity         specificity 
##     "Optimal %LYM:"             "13.13" "0.895652173913044"  "0.16025641025641" 
##                                                             
## "0.282191780821918" "0.806451612903226" "0.498508918617614"

print(c("Optimal n MAC:", exp(MAC.IMI$optimal_cutpoint), MAC.IMI$sensitivity, MAC.IMI$specificity, MAC.IMI$ppv, MAC.IMI$npv, MAC.IMI$AUC))

##                                                 sensitivity         specificity 
##    "Optimal n MAC:"                "21" "0.582608695652174" "0.448717948717949" 
##                                                             
## "0.280334728033473"  "0.74468085106383" "0.498202341137124"

print(c("Optimal %MAC:", PER.MAC.IMI$optimal_cutpoint, PER.MAC.IMI$sensitivity, PER.MAC.IMI$specificity, PER.MAC.IMI$ppv, PER.MAC.IMI$npv, PER.MAC.IMI$AUC))

##                                                 sensitivity         specificity 
##     "Optimal %MAC:"             "36.61" "0.226086956521739" "0.881410256410256" 
##                                                             
## "0.412698412698413" "0.755494505494505" "0.554877369007804"

# CI using bootstrapping

# SCC.IMI
SCC.IMI_optimal_cutpoint <- boot_ci(SCC.IMI, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
SCC.IMI_lower_ci <- subset(SCC.IMI_optimal_cutpoint, SCC.IMI_optimal_cutpoint$quantile == 0.025)$values
exp(SCC.IMI_lower_ci)

## [1] 26

SCC.IMI_upper_ci <- subset(SCC.IMI_optimal_cutpoint, SCC.IMI_optimal_cutpoint$quantile == 0.975)$values
exp(SCC.IMI_upper_ci)

## [1] 246

boot_ci(SCC.IMI, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.268
## 2    0.975  1

boot_ci(SCC.IMI, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile  values
##      <dbl>   <dbl>
## 1    0.025 0.00974
## 2    0.975 0.791

boot_ci(SCC.IMI, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.238
## 2    0.975  0.354

boot_ci(SCC.IMI, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.715
## 2    0.975  1

boot_ci(SCC.IMI, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.422
## 2    0.975  0.546

# NEU.IMI
NEU.IMI_optimal_cutpoint <- boot_ci(NEU.IMI, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
NEU.IMI_lower_ci <- subset(NEU.IMI_optimal_cutpoint, NEU.IMI_optimal_cutpoint$quantile == 0.025)$values
exp(NEU.IMI_lower_ci)

## [1] 25

NEU.IMI_upper_ci <- subset(NEU.IMI_optimal_cutpoint, NEU.IMI_optimal_cutpoint$quantile == 0.975)$values
exp(NEU.IMI_upper_ci)

## [1] 237

boot_ci(NEU.IMI, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.116
## 2    0.975  0.900

boot_ci(NEU.IMI, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.198
## 2    0.975  0.972

boot_ci(NEU.IMI, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.731
## 2    0.975  0.927

boot_ci(NEU.IMI, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.258
## 2    0.975  0.469

boot_ci(NEU.IMI, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.482
## 2    0.975  0.602

# PER.NEU.IMI
PER.NEU.IMI_optimal_cutpoint <- boot_ci(PER.NEU.IMI, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
PER.NEU.IMI_optimal_cutpoint

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025   45.8
## 2    0.975   64.4

PER.NEU.IMI_lower_ci <- subset(PER.NEU.IMI_optimal_cutpoint, PER.NEU.IMI_optimal_cutpoint$quantile == 0.025)$values
PER.NEU.IMI_upper_ci <- subset(PER.NEU.IMI_optimal_cutpoint, PER.NEU.IMI_optimal_cutpoint$quantile == 0.975)$values

boot_ci(PER.NEU.IMI, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.310
## 2    0.975  0.864

boot_ci(PER.NEU.IMI, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.250
## 2    0.975  0.838

boot_ci(PER.NEU.IMI, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.734
## 2    0.975  0.861

boot_ci(PER.NEU.IMI, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.269
## 2    0.975  0.452

boot_ci(PER.NEU.IMI, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.497
## 2    0.975  0.620

# LYM.IMI
LYM.IMI_optimal_cutpoint <- boot_ci(LYM.IMI, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
LYM.IMI_lower_ci <- subset(LYM.IMI_optimal_cutpoint, LYM.IMI_optimal_cutpoint$quantile == 0.025)$values
exp(LYM.IMI_lower_ci)

## [1] 9

LYM.IMI_upper_ci <- subset(LYM.IMI_optimal_cutpoint, LYM.IMI_optimal_cutpoint$quantile == 0.975)$values
exp(LYM.IMI_upper_ci)

## [1] Inf

boot_ci(LYM.IMI, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0    
## 2    0.975  0.930

boot_ci(LYM.IMI, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.104
## 2    0.975  1

boot_ci(LYM.IMI, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.246
## 2    0.975  1

boot_ci(LYM.IMI, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.704
## 2    0.975  0.818

boot_ci(LYM.IMI, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.407
## 2    0.975  0.530

# PER.LYM.IMI
PER.LYM.IMI_optimal_cutpoint <- boot_ci(PER.LYM.IMI, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
PER.LYM.IMI_optimal_cutpoint

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025   12.4
## 2    0.975   27.4

PER.LYM.IMI_lower_ci <- subset(PER.LYM.IMI_optimal_cutpoint, PER.LYM.IMI_optimal_cutpoint$quantile == 0.025)$values
PER.LYM.IMI_upper_ci <- subset(PER.LYM.IMI_optimal_cutpoint, PER.LYM.IMI_optimal_cutpoint$quantile == 0.975)$values

boot_ci(PER.LYM.IMI, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.106
## 2    0.975  0.945

boot_ci(PER.LYM.IMI, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.130
## 2    0.975  0.940

boot_ci(PER.LYM.IMI, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.248
## 2    0.975  0.431

boot_ci(PER.LYM.IMI, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.718
## 2    0.975  0.896

boot_ci(PER.LYM.IMI, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.434
## 2    0.975  0.557

# MAC.IMI
MAC.IMI_optimal_cutpoint <- boot_ci(MAC.IMI, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
MAC.IMI_lower_ci <- subset(MAC.IMI_optimal_cutpoint, MAC.IMI_optimal_cutpoint$quantile == 0.025)$values
exp(MAC.IMI_lower_ci)

## [1] 4

MAC.IMI_upper_ci <- subset(MAC.IMI_optimal_cutpoint, MAC.IMI_optimal_cutpoint$quantile == 0.975)$values
exp(MAC.IMI_upper_ci)

## [1] 64

boot_ci(MAC.IMI, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025    0.2
## 2    0.975    1

boot_ci(MAC.IMI, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025 0.0302
## 2    0.975 0.857

boot_ci(MAC.IMI, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.245
## 2    0.975  0.395

boot_ci(MAC.IMI, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.713
## 2    0.975  1

boot_ci(MAC.IMI, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.437
## 2    0.975  0.558

# PER.MAC.IMI
PER.MAC.IMI_optimal_cutpoint <- boot_ci(PER.MAC.IMI, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
PER.MAC.IMI_optimal_cutpoint

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025   13.4
## 2    0.975   38.3

PER.MAC.IMI_lower_ci <- subset(PER.MAC.IMI_optimal_cutpoint, PER.MAC.IMI_optimal_cutpoint$quantile == 0.025)$values
PER.MAC.IMI_upper_ci <- subset(PER.MAC.IMI_optimal_cutpoint, PER.MAC.IMI_optimal_cutpoint$quantile == 0.975)$values

boot_ci(PER.MAC.IMI, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.178
## 2    0.975  0.878

boot_ci(PER.MAC.IMI, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.236
## 2    0.975  0.929

boot_ci(PER.MAC.IMI, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.277
## 2    0.975  0.537

boot_ci(PER.MAC.IMI, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.724
## 2    0.975  0.849

boot_ci(PER.MAC.IMI, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.492
## 2    0.975  0.617

# Prevalence

# True prevalence

table(Pc.by.cow.C3$IMI.cow)

## 
##   0   1 
## 115 313

prop.table(table(Pc.by.cow.C3$IMI.cow))

## 
##         0         1 
## 0.2686916 0.7313084

#Apparent prevalence

# ---- SCC ----

# optimal cut-off

optimal_scc_C3_IMI_cow <- SCC.IMI$optimal_cutpoint

table(Pc.by.cow.C3$SCC.CO.COW > exp(optimal_scc_C3_IMI_cow))

## 
## FALSE  TRUE 
##   161   267

prop.table(table(Pc.by.cow.C3$SCC.CO.COW > exp(optimal_scc_C3_IMI_cow)))

## 
##     FALSE      TRUE 
## 0.3761682 0.6238318

# confidence intervals

table(Pc.by.cow.C3$SCC.CO.COW > exp(SCC.IMI_lower_ci))

## 
## FALSE  TRUE 
##     6   422

prop.table(table(Pc.by.cow.C3$SCC.CO.COW > exp(SCC.IMI_lower_ci)))

## 
##      FALSE       TRUE 
## 0.01401869 0.98598131

table(Pc.by.cow.C3$SCC.CO.COW > exp(SCC.IMI_upper_ci))

## 
## FALSE  TRUE 
##   328   100

prop.table(table(Pc.by.cow.C3$SCC.CO.COW >exp(SCC.IMI_upper_ci)))

## 
##     FALSE      TRUE 
## 0.7663551 0.2336449

# ---- Neutrophils ----

# optimal cut-off

optimal_neu_C3_IMI_cow <- NEU.IMI$optimal_cutpoint

table(Pc.by.cow.C3$TOTAL.NEU.CO.COW > exp(optimal_neu_C3_IMI_cow))

## 
## FALSE  TRUE 
##   290   140

prop.table(table(Pc.by.cow.C3$TOTAL.NEU.CO.COW > exp(optimal_neu_C3_IMI_cow)))

## 
##     FALSE      TRUE 
## 0.6744186 0.3255814

# confidence intervals

table(Pc.by.cow.C3$TOTAL.NEU.CO.COW > exp(NEU.IMI_lower_ci))

## 
## FALSE  TRUE 
##    54   376

prop.table(table(Pc.by.cow.C3$TOTAL.NEU.CO.COW > exp(NEU.IMI_lower_ci)))

## 
##     FALSE      TRUE 
## 0.1255814 0.8744186

table(Pc.by.cow.C3$TOTAL.NEU.CO.COW > exp(NEU.IMI_upper_ci))

## 
## FALSE  TRUE 
##   387    43

prop.table(table(Pc.by.cow.C3$TOTAL.NEU.CO.COW >exp(NEU.IMI_upper_ci)))

## 
## FALSE  TRUE 
##   0.9   0.1

# ---- Percent neu ----

# Optimal cut-off

optimal_per_neu_C3_IMI_cow <-PER.NEU.IMI$optimal_cutpoint

table(Pc.by.cow.C3$PERCENT.NEU.CO.COW > optimal_per_neu_C3_IMI_cow)

## 
## FALSE  TRUE 
##   259   171

prop.table(table(Pc.by.cow.C3$PERCENT.NEU.CO.COW > optimal_per_neu_C3_IMI_cow))

## 
##     FALSE      TRUE 
## 0.6023256 0.3976744

# confidence intervals

table(Pc.by.cow.C3$PERCENT.NEU.CO.COW > PER.NEU.IMI_lower_ci)

## 
## FALSE  TRUE 
##    81   349

prop.table(table(Pc.by.cow.C3$PERCENT.NEU.CO.COW > PER.NEU.IMI_lower_ci))

## 
##     FALSE      TRUE 
## 0.1883721 0.8116279

table(Pc.by.cow.C3$PERCENT.NEU.CO.COW > PER.NEU.IMI_upper_ci)

## 
## FALSE  TRUE 
##   306   124

prop.table(table(Pc.by.cow.C3$PERCENT.NEU.CO.COW >PER.NEU.IMI_upper_ci))

## 
##     FALSE      TRUE 
## 0.7116279 0.2883721

# ---- Lymphocytes ----

# optimal cut-off

optimal_lym_C3_IMI_cow <-LYM.IMI$optimal_cutpoint

table(Pc.by.cow.C3$TOTAL.LYM.CO.COW > exp(optimal_lym_C3_IMI_cow))

## 
## FALSE  TRUE 
##   226   204

prop.table(table(Pc.by.cow.C3$TOTAL.LYM.CO.COW > exp(optimal_lym_C3_IMI_cow)))

## 
##     FALSE      TRUE 
## 0.5255814 0.4744186

# confidence intervals

table(Pc.by.cow.C3$TOTAL.LYM.CO.COW > exp(LYM.IMI_lower_ci))

## 
## FALSE  TRUE 
##    66   364

prop.table(table(Pc.by.cow.C3$TOTAL.LYM.CO.COW > exp(LYM.IMI_lower_ci)))

## 
##     FALSE      TRUE 
## 0.1534884 0.8465116

table(Pc.by.cow.C3$TOTAL.LYM.CO.COW > exp(LYM.IMI_upper_ci))

## 
## FALSE 
##   430

prop.table(table(Pc.by.cow.C3$TOTAL.LYM.CO.COW >exp(LYM.IMI_upper_ci)))

## 
## FALSE 
##     1

# ---- Percent lym ----

# optimal cut-off

optimal_per_lym_C3_IMI_cow <-PER.LYM.IMI$optimal_cutpoint

table(Pc.by.cow.C3$PERCENT.LYM.CO.COW > optimal_per_lym_C3_IMI_cow)

## 
## FALSE  TRUE 
##    63   367

prop.table(table(Pc.by.cow.C3$PERCENT.LYM.CO.COW > optimal_per_lym_C3_IMI_cow))

## 
##     FALSE      TRUE 
## 0.1465116 0.8534884

# confidence intervals

table(Pc.by.cow.C3$PERCENT.LYM.CO.COW > PER.LYM.IMI_lower_ci)

## 
## FALSE  TRUE 
##    58   372

prop.table(table(Pc.by.cow.C3$PERCENT.LYM.CO.COW > PER.LYM.IMI_lower_ci))

## 
##     FALSE      TRUE 
## 0.1348837 0.8651163

table(Pc.by.cow.C3$PERCENT.LYM.CO.COW > PER.LYM.IMI_upper_ci)

## 
## FALSE  TRUE 
##   394    36

prop.table(table(Pc.by.cow.C3$PERCENT.LYM.CO.COW >PER.LYM.IMI_upper_ci))

## 
##      FALSE       TRUE 
## 0.91627907 0.08372093

# ---- Macrophages ----

# optimal cut-off

optimal_mac_C3_IMI_cow <- MAC.IMI$optimal_cutpoint

table(Pc.by.cow.C3$TOTAL.MAC.CO.COW > exp(optimal_mac_C3_IMI_cow))

## 
## FALSE  TRUE 
##   200   230

prop.table(table(Pc.by.cow.C3$TOTAL.MAC.CO.COW > exp(optimal_mac_C3_IMI_cow)))

## 
##     FALSE      TRUE 
## 0.4651163 0.5348837

# confidence intervals

table(Pc.by.cow.C3$TOTAL.MAC.CO.COW > exp(MAC.IMI_lower_ci))

## 
## FALSE  TRUE 
##    12   418

prop.table(table(Pc.by.cow.C3$TOTAL.MAC.CO.COW > exp(MAC.IMI_lower_ci)))

## 
##      FALSE       TRUE 
## 0.02790698 0.97209302

table(Pc.by.cow.C3$TOTAL.MAC.CO.COW > exp(MAC.IMI_upper_ci))

## 
## FALSE  TRUE 
##   355    75

prop.table(table(Pc.by.cow.C3$TOTAL.MAC.CO.COW >exp(MAC.IMI_upper_ci)))

## 
##     FALSE      TRUE 
## 0.8255814 0.1744186

# Percent MAC

# optimal cut-off

optimal_per_mac_C3_IMI_cow <- PER.MAC.IMI$optimal_cutpoint

table(Pc.by.cow.C3$PERCENT.MAC.CO.COW > optimal_per_mac_C3_IMI_cow)

## 
## FALSE  TRUE 
##   368    62

prop.table(table(Pc.by.cow.C3$PERCENT.MAC.CO.COW > optimal_per_mac_C3_IMI_cow))

## 
##    FALSE     TRUE 
## 0.855814 0.144186

# confidence intervals

table(Pc.by.cow.C3$PERCENT.MAC.CO.COW > PER.MAC.IMI_lower_ci)

## 
## FALSE  TRUE 
##    91   339

prop.table(table(Pc.by.cow.C3$PERCENT.MAC.CO.COW > PER.MAC.IMI_lower_ci))

## 
##     FALSE      TRUE 
## 0.2116279 0.7883721

table(Pc.by.cow.C3$PERCENT.MAC.CO.COW > PER.MAC.IMI_upper_ci)

## 
## FALSE  TRUE 
##   379    51

prop.table(table(Pc.by.cow.C3$PERCENT.MAC.CO.COW >PER.MAC.IMI_upper_ci))

## 
##     FALSE      TRUE 
## 0.8813953 0.1186047

Cow high SCC

# Set a seed for reproducibility

set.seed(123)

# Delaval SCM (SCC>200.000) reference

# n NEU SCM COW
NEU.SCM <- cutpointr(Pc.by.cow.C3, log.TOTAL.NEU.CO.COW, delaval.SCM.CO.COW, 
                            method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(NEU.SCM)

## Method: maximize_metric 
## Predictor: log.TOTAL.NEU.CO.COW 
## Outcome: delaval.SCM.CO.COW 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC   n n_pos n_neg
##  0.9021 428   118   310
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv tp
##            4.6347        1.6466 0.8505      0.7627      0.8839 0.7143 0.9073 90
##  fn fp  tn
##  28 36 274
## 
## Predictor summary: 
##     Data     Min.       5%  1st Qu.   Median     Mean  3rd Qu.      95%
##  Overall 2.197225 2.890372 3.583519 4.127134 4.267632 4.738382 6.126631
##        0 2.197225 2.833213 3.433987 3.871201 3.858973 4.304065 4.895162
##        1 3.367296 4.001777 4.637144 5.043404 5.341228 5.973170 7.852476
##      Max.        SD NAs
##  8.663369 1.0386410   0
##  5.934894 0.6462305   0
##  8.663369 1.1086912   0
## 
## Bootstrap summary: 
##           Variable Min.   5% 1st Qu. Median Mean 3rd Qu.  95% Max.   SD NAs
##   optimal_cutpoint 4.17 4.17    4.42   4.43 4.48    4.63 4.64 4.82 0.17   0
##              AUC_b 0.84 0.87    0.89   0.90 0.90    0.91 0.93 0.94 0.02   0
##            AUC_oob 0.83 0.87    0.89   0.90 0.90    0.92 0.94 0.96 0.02   0
##    sum_sens_spec_b 1.55 1.60    1.64   1.66 1.66    1.68 1.72 1.78 0.04   0
##  sum_sens_spec_oob 1.41 1.52    1.57   1.61 1.61    1.65 1.70 1.79 0.06   0
##              acc_b 0.72 0.76    0.80   0.83 0.83    0.86 0.88 0.90 0.04   0
##            acc_oob 0.67 0.73    0.78   0.82 0.81    0.84 0.87 0.91 0.04   0
##      sensitivity_b 0.66 0.74    0.78   0.83 0.83    0.88 0.95 0.98 0.07   0
##    sensitivity_oob 0.51 0.67    0.73   0.80 0.80    0.86 0.93 1.00 0.08   0
##      specificity_b 0.64 0.70    0.78   0.84 0.83    0.88 0.91 0.95 0.07   0
##    specificity_oob 0.56 0.67    0.76   0.83 0.81    0.88 0.92 0.96 0.08   0
##     cohens_kappa_b 0.43 0.50    0.57   0.61 0.61    0.65 0.70 0.76 0.06   0
##   cohens_kappa_oob 0.33 0.45    0.51   0.57 0.56    0.62 0.68 0.76 0.07   0
##              ppv_b 0.46 0.53    0.60   0.66 0.66    0.72 0.78 0.87 0.08   0
##            ppv_oob 0.38 0.49    0.57   0.64 0.64    0.70 0.78 0.87 0.09   0
##              npv_b 0.85 0.89    0.91   0.93 0.93    0.95 0.97 0.99 0.02   0
##            npv_oob 0.81 0.87    0.90   0.92 0.92    0.94 0.97 1.00 0.03   0

# ROC plot
plot(NEU.SCM)

# Youden index plot
plot_metric(NEU.SCM)

# %NEU SCM COW
PER.NEU.SCM <- cutpointr(Pc.by.cow.C3, PERCENT.NEU.CO.COW, delaval.SCM.CO.COW, 
                            method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(PER.NEU.SCM)

## Method: maximize_metric 
## Predictor: PERCENT.NEU.CO.COW 
## Outcome: delaval.SCM.CO.COW 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC   n n_pos n_neg
##  0.4771 428   118   310
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv tp
##             58.26        1.0302 0.5257      0.4915      0.5387 0.2886 0.7357 58
##  fn  fp  tn
##  60 143 167
## 
## Predictor summary: 
##     Data  Min.      5% 1st Qu. Median     Mean 3rd Qu.    95%  Max.       SD
##  Overall 21.23 36.9360 48.6575 57.405 56.93164 65.3600 75.669 86.98 11.78750
##        0 21.23 37.7180 49.1975 57.270 57.31045 65.2575 75.871 86.98 11.52648
##        1 24.42 34.5775 46.0125 57.675 55.93644 65.4875 75.012 81.00 12.44289
##  NAs
##    0
##    0
##    0
## 
## Bootstrap summary: 
##           Variable  Min.    5% 1st Qu. Median  Mean 3rd Qu.   95% Max.   SD NAs
##   optimal_cutpoint 24.42 30.82   57.20  61.00   Inf   65.35 74.97  Inf  NaN   0
##              AUC_b  0.36  0.42    0.45   0.48  0.48    0.50  0.53 0.59 0.03   0
##            AUC_oob  0.34  0.41    0.45   0.48  0.48    0.51  0.55 0.64 0.04   0
##    sum_sens_spec_b  1.00  1.00    1.03   1.06  1.06    1.09  1.14 1.26 0.04   0
##  sum_sens_spec_oob  0.74  0.87    0.93   0.97  0.97    1.00  1.06 1.16 0.06   0
##              acc_b  0.23  0.29    0.53   0.58  0.56    0.63  0.70 0.77 0.11   0
##            acc_oob  0.20  0.27    0.48   0.54  0.53    0.60  0.67 0.77 0.11   0
##      sensitivity_b  0.00  0.09    0.31   0.44  0.46    0.56  1.00 1.00 0.23   0
##    sensitivity_oob  0.00  0.04    0.23   0.36  0.39    0.49  0.97 1.00 0.23   0
##      specificity_b  0.00  0.01    0.52   0.64  0.60    0.76  0.94 1.00 0.23   0
##    specificity_oob  0.00  0.01    0.49   0.61  0.58    0.73  0.92 1.00 0.23   0
##     cohens_kappa_b  0.00  0.00    0.03   0.05  0.06    0.08  0.13 0.20 0.04   0
##   cohens_kappa_oob -0.20 -0.11   -0.06  -0.03 -0.03    0.00  0.05 0.14 0.05   0
##              ppv_b  0.22  0.26    0.29   0.31  0.32    0.34  0.38 1.00 0.04  27
##            ppv_oob  0.00  0.15    0.22   0.26  0.25    0.28  0.32 0.38 0.06  27
##              npv_b  0.66  0.71    0.73   0.75  0.76    0.77  1.00 1.00 0.07   0
##            npv_oob  0.00  0.50    0.68   0.71  0.69    0.74  0.77 1.00 0.10  63

# ROC plot
plot(PER.NEU.SCM)

## Warning: Removed 27 rows containing non-finite outside the scale range
## (`stat_density()`).

# Youden index plot
plot_metric(PER.NEU.SCM)

# Lym
LYM.SCM  <- cutpointr(Pc.by.cow.C3, log.TOTAL.LYM.CO.COW, delaval.SCM.CO.COW, 
                         method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(LYM.SCM)

## Method: maximize_metric 
## Predictor: log.TOTAL.LYM.CO.COW 
## Outcome: delaval.SCM.CO.COW 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC   n n_pos n_neg
##  0.9203 428   118   310
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv
##            3.4012        1.7063 0.8481      0.8644      0.8419 0.6755 0.9422
##   tp fn fp  tn
##  102 16 49 261
## 
## Predictor summary: 
##     Data     Min.       5%  1st Qu.   Median     Mean  3rd Qu.      95%
##  Overall 0.000000 1.845712 2.544939 3.067782 3.163271 3.610918 4.953354
##        0 0.000000 1.609438 2.302585 2.833213 2.762575 3.208670 3.715714
##        1 2.079442 2.995732 3.533607 4.034201 4.215948 4.755396 6.199995
##      Max.        SD NAs
##  6.921658 0.9821074   0
##  4.521789 0.6315192   0
##  6.921658 0.9611697   0
## 
## Bootstrap summary: 
##           Variable Min.   5% 1st Qu. Median Mean 3rd Qu.  95% Max.   SD NAs
##   optimal_cutpoint 3.14 3.18    3.37   3.40 3.42    3.50 3.53 3.71 0.10   0
##              AUC_b 0.86 0.89    0.91   0.92 0.92    0.93 0.95 0.97 0.02   0
##            AUC_oob 0.85 0.88    0.91   0.92 0.92    0.93 0.95 0.98 0.02   0
##    sum_sens_spec_b 1.59 1.66    1.70   1.72 1.72    1.74 1.78 1.84 0.04   0
##  sum_sens_spec_oob 1.44 1.58    1.64   1.68 1.67    1.71 1.76 1.83 0.05   0
##              acc_b 0.77 0.81    0.84   0.86 0.86    0.88 0.89 0.93 0.03   0
##            acc_oob 0.71 0.79    0.82   0.84 0.84    0.86 0.88 0.92 0.03   0
##      sensitivity_b 0.68 0.78    0.83   0.87 0.86    0.89 0.94 0.99 0.05   0
##    sensitivity_oob 0.52 0.70    0.79   0.83 0.83    0.88 0.93 1.00 0.07   0
##      specificity_b 0.69 0.76    0.83   0.86 0.86    0.89 0.92 0.97 0.04   0
##    specificity_oob 0.65 0.74    0.82   0.85 0.84    0.88 0.92 0.98 0.05   0
##     cohens_kappa_b 0.52 0.58    0.64   0.67 0.67    0.70 0.74 0.83 0.05   0
##   cohens_kappa_oob 0.42 0.53    0.59   0.63 0.63    0.67 0.72 0.82 0.06   0
##              ppv_b 0.52 0.59    0.66   0.70 0.70    0.74 0.80 0.91 0.06   0
##            ppv_oob 0.45 0.56    0.63   0.68 0.68    0.72 0.79 0.92 0.07   0
##              npv_b 0.88 0.91    0.93   0.94 0.94    0.96 0.97 1.00 0.02   0
##            npv_oob 0.83 0.88    0.91   0.93 0.93    0.95 0.97 1.00 0.03   0

# ROC plot
plot(LYM.SCM)

# Youden index plot
plot_metric(LYM.SCM)

# %LYM SCM COW
PER.LYM.SCM <- cutpointr(Pc.by.cow.C3, PERCENT.LYM.CO.COW, delaval.SCM.CO.COW, 
                            method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 0

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(PER.LYM.SCM)

## Method: maximize_metric 
## Predictor: PERCENT.LYM.CO.COW 
## Outcome: delaval.SCM.CO.COW 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##    AUC   n n_pos n_neg
##  0.571 428   310   118
## 
##  optimal_cutpoint sum_sens_spec   acc sensitivity specificity    ppv   npv tp
##             22.45        1.1922 0.472      0.3194      0.8729 0.8684 0.328 99
##   fn fp  tn
##  211 15 103
## 
## Predictor summary: 
##     Data Min.     5% 1st Qu. Median     Mean 3rd Qu.     95% Max.       SD NAs
##  Overall 5.08 10.147 15.6800 19.205 19.32189 22.6625 29.3785 51.1 5.879269   0
##        0 5.66 10.149 15.6750 19.525 19.69710 23.4225 29.5805 51.1 6.163532   0
##        1 5.08 11.199 15.7225 17.990 18.33619 21.0925 27.4830 36.5 4.946434   0
## 
## Bootstrap summary: 
##           Variable  Min.    5% 1st Qu. Median  Mean 3rd Qu.   95%  Max.   SD
##   optimal_cutpoint 16.49 18.44   19.06  22.45 21.30   22.45 22.73 24.44 1.63
##              AUC_b  0.48  0.52    0.55   0.57  0.57    0.59  0.62  0.65 0.03
##            AUC_oob  0.44  0.51    0.55   0.57  0.57    0.60  0.64  0.71 0.04
##    sum_sens_spec_b  1.06  1.14    1.18   1.21  1.21    1.23  1.27  1.33 0.04
##  sum_sens_spec_oob  0.85  1.02    1.10   1.15  1.14    1.19  1.25  1.34 0.07
##              acc_b  0.36  0.44    0.47   0.50  0.52    0.57  0.62  0.67 0.06
##            acc_oob  0.33  0.41    0.45   0.48  0.49    0.52  0.58  0.65 0.05
##      sensitivity_b  0.18  0.28    0.31   0.35  0.41    0.53  0.62  0.77 0.12
##    sensitivity_oob  0.16  0.25    0.30   0.34  0.39    0.50  0.60  0.69 0.12
##      specificity_b  0.43  0.59    0.69   0.85  0.80    0.89  0.92  0.97 0.12
##    specificity_oob  0.22  0.48    0.61   0.82  0.75    0.88  0.92  1.00 0.15
##     cohens_kappa_b  0.03  0.09    0.12   0.14  0.15    0.17  0.21  0.28 0.04
##   cohens_kappa_oob -0.14  0.02    0.07   0.10  0.10    0.13  0.17  0.25 0.05
##              ppv_b  0.72  0.78    0.82   0.85  0.85    0.88  0.92  0.97 0.04
##            ppv_oob  0.66  0.72    0.77   0.82  0.82    0.87  0.92  1.00 0.06
##              npv_b  0.25  0.29    0.32   0.34  0.34    0.36  0.40  0.45 0.03
##            npv_oob  0.19  0.26    0.29   0.32  0.32    0.34  0.38  0.46 0.04
##  NAs
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0

# ROC plot
plot(PER.LYM.SCM)

# Youden index plot
plot_metric(PER.LYM.SCM)

# MAC SCM COW
MAC.SCM <- cutpointr(Pc.by.cow.C3, log.TOTAL.MAC.CO.COW, delaval.SCM.CO.COW, 
                        method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(MAC.SCM)

## Method: maximize_metric 
## Predictor: log.TOTAL.MAC.CO.COW 
## Outcome: delaval.SCM.CO.COW 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC   n n_pos n_neg
##  0.9122 428   118   310
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv tp
##            3.5835         1.666 0.8341      0.8305      0.8355 0.6577 0.9283 98
##  fn fp  tn
##  20 51 259
## 
## Predictor summary: 
##     Data     Min.       5%  1st Qu.   Median     Mean  3rd Qu.      95%
##  Overall 0.000000 1.791759 2.564949 3.178054 3.291112 3.871201 5.383950
##        0 0.000000 1.609438 2.397895 2.833213 2.846021 3.392722 3.970292
##        1 2.833213 3.029510 3.719596 4.241301 4.460422 5.055237 6.696229
##      Max.       SD NAs
##  8.318254 1.121363   0
##  4.663439 0.742693   0
##  8.318254 1.108391   0
## 
## Bootstrap summary: 
##           Variable Min.   5% 1st Qu. Median Mean 3rd Qu.  95% Max.   SD NAs
##   optimal_cutpoint 3.04 3.58    3.58   3.58 3.63    3.64 3.87 4.06 0.11   0
##              AUC_b 0.86 0.89    0.90   0.91 0.91    0.92 0.93 0.96 0.01   0
##            AUC_oob 0.84 0.88    0.90   0.91 0.91    0.93 0.94 0.97 0.02   0
##    sum_sens_spec_b 1.53 1.61    1.64   1.67 1.67    1.70 1.73 1.82 0.04   0
##  sum_sens_spec_oob 1.41 1.53    1.60   1.64 1.64    1.68 1.74 1.82 0.06   0
##              acc_b 0.69 0.81    0.83   0.84 0.84    0.86 0.88 0.92 0.02   0
##            acc_oob 0.67 0.79    0.82   0.83 0.83    0.85 0.87 0.91 0.02   0
##      sensitivity_b 0.62 0.73    0.79   0.82 0.82    0.85 0.88 0.95 0.05   0
##    sensitivity_oob 0.45 0.63    0.76   0.81 0.79    0.84 0.90 1.00 0.08   0
##      specificity_b 0.60 0.80    0.83   0.85 0.85    0.87 0.93 0.99 0.04   0
##    specificity_oob 0.56 0.78    0.82   0.84 0.85    0.87 0.93 1.00 0.04   0
##     cohens_kappa_b 0.40 0.56    0.60   0.63 0.63    0.66 0.70 0.79 0.04   0
##   cohens_kappa_oob 0.36 0.51    0.57   0.60 0.60    0.64 0.68 0.78 0.05   0
##              ppv_b 0.45 0.61    0.65   0.68 0.69    0.71 0.80 0.96 0.06   0
##            ppv_oob 0.43 0.57    0.62   0.66 0.67    0.70 0.79 1.00 0.07   0
##              npv_b 0.85 0.89    0.91   0.93 0.92    0.94 0.95 0.97 0.02   0
##            npv_oob 0.77 0.86    0.90   0.92 0.92    0.94 0.96 1.00 0.03   0

# ROC plot
plot(MAC.SCM)

# Youden index plot
plot_metric(MAC.SCM)

# %MAC SCM COW
PER.MAC.SCM <- cutpointr(Pc.by.cow.C3, PERCENT.MAC.CO.COW, delaval.SCM.CO.COW, 
                            method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(PER.MAC.SCM)

## Method: maximize_metric 
## Predictor: PERCENT.MAC.CO.COW 
## Outcome: delaval.SCM.CO.COW 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC   n n_pos n_neg
##  0.5656 428   118   310
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv   npv tp
##              23.3        1.1976 0.6051      0.5847      0.6129 0.3651 0.795 69
##  fn  fp  tn
##  49 120 190
## 
## Predictor summary: 
##     Data Min.     5% 1st Qu. Median     Mean 3rd Qu.     95%  Max.       SD NAs
##  Overall 3.51 8.6400 15.2325 21.955 23.74565 30.8125 44.7650 73.11 11.55696   0
##        0 3.51 8.7075 15.1350 21.155 22.99168 29.8275 42.3935 73.11 11.10716   0
##        1 4.67 8.6385 15.6950 24.695 25.72644 33.3975 49.4035 59.41 12.49658   0
## 
## Bootstrap summary: 
##           Variable  Min.    5% 1st Qu. Median  Mean 3rd Qu.   95%  Max.   SD
##   optimal_cutpoint 10.73 23.14   23.14  23.30 23.58   23.47 24.27 45.43 2.19
##              AUC_b  0.44  0.51    0.54   0.56  0.56    0.59  0.62  0.68 0.03
##            AUC_oob  0.42  0.49    0.54   0.57  0.57    0.60  0.64  0.72 0.04
##    sum_sens_spec_b  1.02  1.12    1.17   1.20  1.20    1.24  1.29  1.39 0.05
##  sum_sens_spec_oob  0.91  1.04    1.12   1.17  1.17    1.22  1.29  1.39 0.08
##              acc_b  0.34  0.57    0.59   0.61  0.61    0.63  0.65  0.76 0.03
##            acc_oob  0.32  0.54    0.58   0.60  0.60    0.62  0.65  0.72 0.04
##      sensitivity_b  0.09  0.50    0.56   0.59  0.59    0.62  0.67  0.96 0.07
##    sensitivity_oob  0.02  0.43    0.51   0.56  0.56    0.61  0.67  0.93 0.09
##      specificity_b  0.14  0.56    0.59   0.62  0.62    0.64  0.67  0.96 0.06
##    specificity_oob  0.10  0.54    0.58   0.61  0.61    0.64  0.68  0.96 0.07
##     cohens_kappa_b  0.03  0.10    0.14   0.17  0.17    0.20  0.25  0.33 0.05
##   cohens_kappa_oob -0.07  0.03    0.10   0.14  0.14    0.18  0.24  0.34 0.06
##              ppv_b  0.25  0.31    0.35   0.37  0.37    0.39  0.43  0.68 0.04
##            ppv_oob  0.17  0.27    0.32   0.35  0.35    0.38  0.43  0.56 0.05
##              npv_b  0.70  0.75    0.78   0.80  0.80    0.81  0.84  0.94 0.03
##            npv_oob  0.57  0.72    0.76   0.79  0.78    0.81  0.84  0.91 0.04
##  NAs
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0

# ROC plot
plot(PER.MAC.SCM)

# Youden index plot
plot_metric(PER.MAC.SCM)

# Summary of results when using SCM as reference

print(c("Optimal n NEU:", exp(NEU.SCM$optimal_cutpoint), NEU.SCM$sensitivity, NEU.SCM$specificity, NEU.SCM$ppv, NEU.SCM$npv, NEU.SCM$AUC))

##                                                 sensitivity         specificity 
##    "Optimal n NEU:"               "103"  "0.76271186440678" "0.883870967741936" 
##                                                             
## "0.714285714285714" "0.907284768211921" "0.902104975396391"

print(c("Optimal %NEU:", PER.NEU.SCM$optimal_cutpoint, PER.NEU.SCM$sensitivity, PER.NEU.SCM$specificity, PER.NEU.SCM$ppv, PER.NEU.SCM$npv, PER.NEU.SCM$AUC))

##                                                 sensitivity         specificity 
##     "Optimal %NEU:"             "58.26" "0.491525423728814" "0.538709677419355" 
##                                                             
## "0.288557213930348"  "0.73568281938326" "0.477104975396391"

print(c("Optimal n LYM:", exp(LYM.SCM$optimal_cutpoint), LYM.SCM$sensitivity, LYM.SCM$specificity, LYM.SCM$ppv, LYM.SCM$npv, LYM.SCM$AUC))

##                                                 sensitivity         specificity 
##    "Optimal n LYM:"                "30" "0.864406779661017" "0.841935483870968" 
##                                                             
## "0.675496688741722" "0.942238267148014" "0.920270639693822"

print(c("Optimal %LYM:", PER.LYM.SCM$optimal_cutpoint, PER.LYM.SCM$sensitivity, PER.LYM.SCM$specificity, PER.LYM.SCM$ppv, PER.LYM.SCM$npv, PER.LYM.SCM$AUC))

##                                                 sensitivity         specificity 
##     "Optimal %LYM:"             "22.45" "0.319354838709677" "0.872881355932203" 
##                                                             
## "0.868421052631579" "0.328025477707006" "0.570995079278294"

print(c("Optimal n MAC:", exp(MAC.SCM$optimal_cutpoint), MAC.SCM$sensitivity, MAC.SCM$specificity, MAC.SCM$ppv, MAC.SCM$npv, MAC.SCM$AUC))

##                                                 sensitivity         specificity 
##    "Optimal n MAC:"                "36" "0.830508474576271" "0.835483870967742" 
##                                                             
## "0.657718120805369"  "0.92831541218638" "0.912178786221979"

print(c("Optimal %MAC:", PER.MAC.SCM$optimal_cutpoint, PER.MAC.SCM$sensitivity, PER.MAC.SCM$specificity, PER.MAC.SCM$ppv, PER.MAC.SCM$npv, PER.MAC.SCM$AUC))

##                                                 sensitivity         specificity 
##     "Optimal %MAC:"              "23.3" "0.584745762711864" "0.612903225806452" 
##                                                             
## "0.365079365079365" "0.794979079497908" "0.565554948059049"

# NEU.SCM
NEU.SCM_optimal_cutpoint <- boot_ci(NEU.SCM, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
NEU.SCM_lower_ci <- subset(NEU.SCM_optimal_cutpoint, NEU.SCM_optimal_cutpoint$quantile == 0.025)$values
exp(NEU.SCM_lower_ci)

## [1] 65

NEU.SCM_upper_ci <- subset(NEU.SCM_optimal_cutpoint, NEU.SCM_optimal_cutpoint$quantile == 0.975)$values
exp(NEU.SCM_upper_ci)

## [1] 106

boot_ci(NEU.SCM, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.722
## 2    0.975  0.958

boot_ci(NEU.SCM, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.683
## 2    0.975  0.917

boot_ci(NEU.SCM, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.510
## 2    0.975  0.793

boot_ci(NEU.SCM, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.888
## 2    0.975  0.978

boot_ci(NEU.SCM, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.870
## 2    0.975  0.931

# PER.NEU.SCM
PER.NEU.SCM_optimal_cutpoint <- boot_ci(PER.NEU.SCM, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
PER.NEU.SCM_optimal_cutpoint

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025   24.4
## 2    0.975   75.3

PER.NEU.SCM_lower_ci <- subset(PER.NEU.SCM_optimal_cutpoint, PER.NEU.SCM_optimal_cutpoint$quantile == 0.025)$values
PER.NEU.SCM_upper_ci <- subset(PER.NEU.SCM_optimal_cutpoint, PER.NEU.SCM_optimal_cutpoint$quantile == 0.975)$values

boot_ci(PER.NEU.SCM, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025 0.0588
## 2    0.975 1

boot_ci(PER.NEU.SCM, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile  values
##      <dbl>   <dbl>
## 1    0.025 0.00627
## 2    0.975 0.962

boot_ci(PER.NEU.SCM, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.255
## 2    0.975  0.409

boot_ci(PER.NEU.SCM, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.699
## 2    0.975  1

boot_ci(PER.NEU.SCM, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.414
## 2    0.975  0.542

# LYM.SCM
LYM.SCM_optimal_cutpoint <- boot_ci(LYM.SCM, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
LYM.SCM_lower_ci <- subset(LYM.SCM_optimal_cutpoint, LYM.SCM_optimal_cutpoint$quantile == 0.025)$values
exp(LYM.SCM_lower_ci)

## [1] 24

LYM.SCM_upper_ci <- subset(LYM.SCM_optimal_cutpoint, LYM.SCM_optimal_cutpoint$quantile == 0.975)$values
exp(LYM.SCM_upper_ci)

## [1] 38

boot_ci(LYM.SCM, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.767
## 2    0.975  0.951

boot_ci(LYM.SCM, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.748
## 2    0.975  0.936

boot_ci(LYM.SCM, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.572
## 2    0.975  0.829

boot_ci(LYM.SCM, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.909
## 2    0.975  0.977

boot_ci(LYM.SCM, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.888
## 2    0.975  0.948

# PER.LYM.SCM
PER.LYM.SCM_optimal_cutpoint <- boot_ci(PER.LYM.SCM, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
PER.LYM.SCM_optimal_cutpoint

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025   18.4
## 2    0.975   23.3

PER.LYM.SCM_lower_ci <- subset(PER.LYM.SCM_optimal_cutpoint, PER.LYM.SCM_optimal_cutpoint$quantile == 0.025)$values
PER.LYM.SCM_upper_ci <- subset(PER.LYM.SCM_optimal_cutpoint, PER.LYM.SCM_optimal_cutpoint$quantile == 0.975)$values

boot_ci(PER.LYM.SCM, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.267
## 2    0.975  0.630

boot_ci(PER.LYM.SCM, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.565
## 2    0.975  0.930

boot_ci(PER.LYM.SCM, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.771
## 2    0.975  0.926

boot_ci(PER.LYM.SCM, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.284
## 2    0.975  0.405

boot_ci(PER.LYM.SCM, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.513
## 2    0.975  0.625

# MAC.SCM
MAC.SCM_optimal_cutpoint <- boot_ci(MAC.SCM, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
MAC.SCM_lower_ci <- subset(MAC.SCM_optimal_cutpoint, MAC.SCM_optimal_cutpoint$quantile == 0.025)$values
exp(MAC.SCM_lower_ci)

## [1] 32

MAC.SCM_upper_ci <- subset(MAC.SCM_optimal_cutpoint, MAC.SCM_optimal_cutpoint$quantile == 0.975)$values
exp(MAC.SCM_upper_ci)

## [1] 52

boot_ci(MAC.SCM, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.702
## 2    0.975  0.894

boot_ci(MAC.SCM, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.793
## 2    0.975  0.947

boot_ci(MAC.SCM, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.588
## 2    0.975  0.837

boot_ci(MAC.SCM, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.886
## 2    0.975  0.957

boot_ci(MAC.SCM, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.882
## 2    0.975  0.940

# PER.MAC.SCM
PER.MAC.SCM_optimal_cutpoint <- boot_ci(PER.MAC.SCM, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
PER.MAC.SCM_optimal_cutpoint

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025   22.0
## 2    0.975   26.6

PER.MAC.SCM_lower_ci <- subset(PER.MAC.SCM_optimal_cutpoint, PER.MAC.SCM_optimal_cutpoint$quantile == 0.025)$values
PER.MAC.SCM_upper_ci <- subset(PER.MAC.SCM_optimal_cutpoint, PER.MAC.SCM_optimal_cutpoint$quantile == 0.975)$values

boot_ci(PER.MAC.SCM, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.474
## 2    0.975  0.695

boot_ci(PER.MAC.SCM, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.543
## 2    0.975  0.692

boot_ci(PER.MAC.SCM, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.298
## 2    0.975  0.449

boot_ci(PER.MAC.SCM, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.744
## 2    0.975  0.848

boot_ci(PER.MAC.SCM, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.502
## 2    0.975  0.629

# Prevalence

# True prevalence

table(Pc.by.cow.C3$delaval.SCM.CO.COW)

## 
##   0   1 
## 310 118

prop.table(table(Pc.by.cow.C3$delaval.SCM.CO.COW))

## 
##         0         1 
## 0.7242991 0.2757009

#Apparent prevalence

# ---- Neutrophils ----

# optimal cut-off

optimal_neu_C3_SCM_cow <- NEU.SCM$optimal_cutpoint

table(Pc.by.cow.C3$TOTAL.NEU.CO.COW > exp(optimal_neu_C3_SCM_cow))

## 
## FALSE  TRUE 
##   307   123

prop.table(table(Pc.by.cow.C3$TOTAL.NEU.CO.COW > exp(optimal_neu_C3_SCM_cow)))

## 
##     FALSE      TRUE 
## 0.7139535 0.2860465

# confidence intervals

table(Pc.by.cow.C3$TOTAL.NEU.CO.COW > exp(NEU.SCM_lower_ci))

## 
## FALSE  TRUE 
##   223   207

prop.table(table(Pc.by.cow.C3$TOTAL.NEU.CO.COW > exp(NEU.SCM_lower_ci)))

## 
##     FALSE      TRUE 
## 0.5186047 0.4813953

table(Pc.by.cow.C3$TOTAL.NEU.CO.COW > exp(NEU.SCM_upper_ci))

## 
## FALSE  TRUE 
##   309   121

prop.table(table(Pc.by.cow.C3$TOTAL.NEU.CO.COW >exp(NEU.SCM_upper_ci)))

## 
##     FALSE      TRUE 
## 0.7186047 0.2813953

# ---- Percent neu ----

# Optimal cut-off

optimal_per_neu_C3_SCM_cow <-PER.NEU.SCM$optimal_cutpoint

table(Pc.by.cow.C3$PERCENT.NEU.CO.COW > optimal_per_neu_C3_SCM_cow)

## 
## FALSE  TRUE 
##   228   202

prop.table(table(Pc.by.cow.C3$PERCENT.NEU.CO.COW > optimal_per_neu_C3_SCM_cow))

## 
##     FALSE      TRUE 
## 0.5302326 0.4697674

# confidence intervals

table(Pc.by.cow.C3$PERCENT.NEU.CO.COW > PER.NEU.SCM_lower_ci)

## 
## FALSE  TRUE 
##     2   428

prop.table(table(Pc.by.cow.C3$PERCENT.NEU.CO.COW > PER.NEU.SCM_lower_ci))

## 
##       FALSE        TRUE 
## 0.004651163 0.995348837

table(Pc.by.cow.C3$PERCENT.NEU.CO.COW > PER.NEU.SCM_upper_ci)

## 
## FALSE  TRUE 
##   406    24

prop.table(table(Pc.by.cow.C3$PERCENT.NEU.CO.COW >PER.NEU.SCM_upper_ci))

## 
##      FALSE       TRUE 
## 0.94418605 0.05581395

# ---- Lymphocytes ----

# optimal cut-off

optimal_lym_C3_SCM_cow <-LYM.SCM$optimal_cutpoint

table(Pc.by.cow.C3$TOTAL.LYM.CO.COW > exp(optimal_lym_C3_SCM_cow))

## 
## FALSE  TRUE 
##   284   146

prop.table(table(Pc.by.cow.C3$TOTAL.LYM.CO.COW > exp(optimal_lym_C3_SCM_cow)))

## 
##     FALSE      TRUE 
## 0.6604651 0.3395349

# confidence intervals

table(Pc.by.cow.C3$TOTAL.LYM.CO.COW > exp(LYM.SCM_lower_ci))

## 
## FALSE  TRUE 
##   243   187

prop.table(table(Pc.by.cow.C3$TOTAL.LYM.CO.COW > exp(LYM.SCM_lower_ci)))

## 
##     FALSE      TRUE 
## 0.5651163 0.4348837

table(Pc.by.cow.C3$TOTAL.LYM.CO.COW > exp(LYM.SCM_upper_ci))

## 
## FALSE  TRUE 
##   326   104

prop.table(table(Pc.by.cow.C3$TOTAL.LYM.CO.COW >exp(LYM.SCM_upper_ci)))

## 
##     FALSE      TRUE 
## 0.7581395 0.2418605

# ---- Percent lym ----

# optimal cut-off

optimal_per_lym_C3_SCM_cow <-PER.LYM.SCM$optimal_cutpoint

table(Pc.by.cow.C3$PERCENT.LYM.CO.COW > optimal_per_lym_C3_SCM_cow)

## 
## FALSE  TRUE 
##   317   113

prop.table(table(Pc.by.cow.C3$PERCENT.LYM.CO.COW > optimal_per_lym_C3_SCM_cow))

## 
##     FALSE      TRUE 
## 0.7372093 0.2627907

# confidence intervals

table(Pc.by.cow.C3$PERCENT.LYM.CO.COW > PER.LYM.SCM_lower_ci)

## 
## FALSE  TRUE 
##   189   241

prop.table(table(Pc.by.cow.C3$PERCENT.LYM.CO.COW > PER.LYM.SCM_lower_ci))

## 
##     FALSE      TRUE 
## 0.4395349 0.5604651

table(Pc.by.cow.C3$PERCENT.LYM.CO.COW > PER.LYM.SCM_upper_ci)

## 
## FALSE  TRUE 
##   335    95

prop.table(table(Pc.by.cow.C3$PERCENT.LYM.CO.COW >PER.LYM.SCM_upper_ci))

## 
##     FALSE      TRUE 
## 0.7790698 0.2209302

# ---- Macrophages ----

# optimal cut-off

optimal_mac_C3_SCM_cow <- MAC.SCM$optimal_cutpoint

table(Pc.by.cow.C3$TOTAL.MAC.CO.COW > exp(optimal_mac_C3_SCM_cow))

## 
## FALSE  TRUE 
##   286   144

prop.table(table(Pc.by.cow.C3$TOTAL.MAC.CO.COW > exp(optimal_mac_C3_SCM_cow)))

## 
##     FALSE      TRUE 
## 0.6651163 0.3348837

# confidence intervals

table(Pc.by.cow.C3$TOTAL.MAC.CO.COW > exp(MAC.SCM_lower_ci))

## 
## FALSE  TRUE 
##   270   160

prop.table(table(Pc.by.cow.C3$TOTAL.MAC.CO.COW > exp(MAC.SCM_lower_ci)))

## 
##    FALSE     TRUE 
## 0.627907 0.372093

table(Pc.by.cow.C3$TOTAL.MAC.CO.COW > exp(MAC.SCM_upper_ci))

## 
## FALSE  TRUE 
##   335    95

prop.table(table(Pc.by.cow.C3$TOTAL.MAC.CO.COW >exp(MAC.SCM_upper_ci)))

## 
##     FALSE      TRUE 
## 0.7790698 0.2209302

# Percent MAC

# optimal cut-off

optimal_per_mac_C3_SCM_cow <- PER.MAC.SCM$optimal_cutpoint

table(Pc.by.cow.C3$PERCENT.MAC.CO.COW > optimal_per_mac_C3_SCM_cow)

## 
## FALSE  TRUE 
##   241   189

prop.table(table(Pc.by.cow.C3$PERCENT.MAC.CO.COW > optimal_per_mac_C3_SCM_cow))

## 
##     FALSE      TRUE 
## 0.5604651 0.4395349

# confidence intervals

table(Pc.by.cow.C3$PERCENT.MAC.CO.COW > PER.MAC.SCM_lower_ci)

## 
## FALSE  TRUE 
##   216   214

prop.table(table(Pc.by.cow.C3$PERCENT.MAC.CO.COW > PER.MAC.SCM_lower_ci))

## 
##     FALSE      TRUE 
## 0.5023256 0.4976744

table(Pc.by.cow.C3$PERCENT.MAC.CO.COW > PER.MAC.SCM_upper_ci)

## 
## FALSE  TRUE 
##   280   150

prop.table(table(Pc.by.cow.C3$PERCENT.MAC.CO.COW >PER.MAC.SCM_upper_ci))

## 
##     FALSE      TRUE 
## 0.6511628 0.3488372

Quarter IMI

# Set a seed for reproducibility

set.seed(123)

# IMI quarter referent

# SCC IMI cow
SCC.IMI <- cutpointr(Pc.by.quarter.C3, log.SCC, IMI, 
                     method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 0

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(SCC.IMI)

## Method: maximize_metric 
## Predictor: log.SCC 
## Outcome: IMI 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC    n n_pos n_neg
##  0.4941 1590  1046   544
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv
##            3.7612        1.0279 0.6396       0.912      0.1158 0.6648 0.4065
##   tp fn  fp tn
##  954 92 481 63
## 
## Predictor summary: 
##     Data     Min.       5%  1st Qu.   Median     Mean  3rd Qu.      95%
##  Overall 1.791759 3.526361 4.248495 4.709530 4.782004 5.214936 6.217703
##        0 1.945910 3.533607 4.248495 4.718499 4.757269 5.185985 6.098074
##        1 1.791759 3.401197 4.244898 4.709530 4.829564 5.279387 6.878855
##      Max.        SD NAs
##  8.660081 0.8996189   0
##  8.326275 0.8273682   0
##  8.660081 1.0234700   0
## 
## Bootstrap summary: 
##           Variable  Min.    5% 1st Qu. Median Mean 3rd Qu.  95% Max.   SD NAs
##   optimal_cutpoint  1.95  3.30    3.71   3.76 4.04    4.64 4.88 5.66 0.53   0
##              AUC_b  0.44  0.47    0.48   0.49 0.49    0.50 0.52 0.55 0.02   0
##            AUC_oob  0.43  0.46    0.48   0.50 0.49    0.51 0.53 0.56 0.02   0
##    sum_sens_spec_b  1.00  1.02    1.03   1.04 1.04    1.05 1.07 1.12 0.02   0
##  sum_sens_spec_oob  0.87  0.94    0.98   1.00 1.00    1.02 1.04 1.08 0.03   0
##              acc_b  0.37  0.49    0.54   0.63 0.60    0.65 0.67 0.70 0.06   0
##            acc_oob  0.36  0.45    0.52   0.62 0.58    0.64 0.66 0.70 0.07   0
##      sensitivity_b  0.12  0.42    0.58   0.91 0.78    0.93 0.99 1.00 0.20   0
##    sensitivity_oob  0.11  0.39    0.56   0.89 0.77    0.92 0.98 1.00 0.21   0
##      specificity_b  0.00  0.04    0.11   0.14 0.26    0.47 0.63 0.90 0.20   0
##    specificity_oob  0.00  0.03    0.09   0.13 0.23    0.42 0.58 0.82 0.19   0
##     cohens_kappa_b  0.00  0.02    0.03   0.04 0.04    0.05 0.07 0.11 0.02   0
##   cohens_kappa_oob -0.10 -0.06   -0.02   0.00 0.00    0.02 0.04 0.09 0.03   0
##              ppv_b  0.63  0.65    0.66   0.67 0.67    0.68 0.69 0.73 0.01   0
##            ppv_oob  0.53  0.62    0.64   0.66 0.66    0.67 0.68 0.72 0.02   0
##              npv_b  0.32  0.35    0.37   0.40 0.42    0.45 0.59 1.00 0.08   0
##            npv_oob  0.17  0.29    0.32   0.35 0.37    0.40 0.53 0.80 0.08   2

# ROC plot
plot(SCC.IMI)

# Youden index plot
plot_metric(SCC.IMI)

# NEU IMI cow
NEU.IMI <- cutpointr(Pc.by.quarter.C3, log.Neu, IMI, 
                     method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(NEU.IMI)

## Method: maximize_metric 
## Predictor: log.Neu 
## Outcome: IMI 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC    n n_pos n_neg
##  0.5278 1594   544  1050
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv npv  tp
##            3.6376        1.0518 0.4542      0.7518         0.3 0.3575 0.7 409
##   fn  fp  tn
##  135 735 315
## 
## Predictor summary: 
##     Data     Min.       5%  1st Qu.   Median     Mean  3rd Qu.      95%
##  Overall 1.609438 2.833213 3.555348 4.025352 4.128486 4.584967 5.652652
##        0 1.609438 2.833213 3.526361 4.007333 4.083754 4.564348 5.497954
##        1 1.609438 2.833213 3.637586 4.077537 4.214824 4.656330 6.099611
##      Max.        SD NAs
##  9.429396 0.9611833   0
##  8.848940 0.8999282   0
##  9.429396 1.0651172   0
## 
## Bootstrap summary: 
##           Variable  Min.    5% 1st Qu. Median Mean 3rd Qu.  95% Max.   SD NAs
##   optimal_cutpoint  2.77  3.61    3.64   4.09 4.25    4.64 5.38 7.97 0.64   0
##              AUC_b  0.47  0.50    0.52   0.53 0.53    0.54 0.55 0.58 0.02   0
##            AUC_oob  0.46  0.49    0.51   0.53 0.53    0.54 0.56 0.59 0.02   0
##    sum_sens_spec_b  1.02  1.04    1.05   1.07 1.07    1.08 1.10 1.14 0.02   0
##  sum_sens_spec_oob  0.89  0.97    1.00   1.02 1.02    1.04 1.06 1.11 0.03   0
##              acc_b  0.36  0.45    0.47   0.54 0.55    0.62 0.66 0.69 0.08   0
##            acc_oob  0.33  0.42    0.45   0.52 0.52    0.59 0.65 0.69 0.08   0
##      sensitivity_b  0.03  0.10    0.27   0.51 0.49    0.75 0.79 0.99 0.25   0
##    sensitivity_oob  0.01  0.08    0.23   0.47 0.46    0.73 0.77 0.94 0.25   0
##      specificity_b  0.05  0.28    0.31   0.56 0.57    0.79 0.95 1.00 0.24   0
##    specificity_oob  0.03  0.26    0.30   0.54 0.56    0.77 0.94 0.99 0.24   0
##     cohens_kappa_b  0.02  0.03    0.05   0.06 0.06    0.07 0.10 0.13 0.02   0
##   cohens_kappa_oob -0.10 -0.03    0.00   0.02 0.01    0.03 0.06 0.09 0.03   0
##              ppv_b  0.32  0.35    0.36   0.38 0.40    0.41 0.51 0.93 0.05   0
##            ppv_oob  0.23  0.31    0.34   0.35 0.36    0.37 0.45 0.59 0.04   0
##              npv_b  0.64  0.66    0.68   0.69 0.69    0.71 0.73 0.86 0.02   0
##            npv_oob  0.48  0.63    0.65   0.67 0.67    0.68 0.71 0.75 0.02   0

# ROC plot
plot(NEU.IMI)

# Youden index plot
plot_metric(NEU.IMI)

# %NEU IMI cow
PER.NEU.IMI <- cutpointr(Pc.by.quarter.C3, PERCENT01_N, IMI, 
                         method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(PER.NEU.IMI)

## Method: maximize_metric 
## Predictor: PERCENT01_N 
## Outcome: IMI 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC    n n_pos n_neg
##  0.5265 1594   544  1050
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv
##             59.65        1.0575 0.5245      0.5423      0.5152 0.3669 0.6848
##   tp  fn  fp  tn
##  295 249 509 541
## 
## Predictor summary: 
##     Data  Min.     5% 1st Qu. Median     Mean 3rd Qu.     95%  Max.       SD
##  Overall 20.16 38.920 51.3925 59.700 59.01367 67.5150 76.6035 87.31 11.49611
##        0 20.16 37.931 50.7325 59.190 58.64175 67.2550 76.2700 87.14 11.66787
##        1 26.92 39.793 52.8600 60.525 59.73153 68.0525 76.7355 87.31 11.13256
##  NAs
##    0
##    0
##    0
## 
## Bootstrap summary: 
##           Variable  Min.    5% 1st Qu. Median  Mean 3rd Qu.   95%  Max.   SD
##   optimal_cutpoint 32.91 48.94   52.88  57.84 56.78   59.70 63.45 76.67 4.98
##              AUC_b  0.47  0.50    0.52   0.53  0.53    0.54  0.55  0.57 0.01
##            AUC_oob  0.46  0.50    0.51   0.53  0.53    0.54  0.56  0.59 0.02
##    sum_sens_spec_b  1.01  1.04    1.06   1.07  1.07    1.09  1.11  1.16 0.02
##  sum_sens_spec_oob  0.91  0.97    1.01   1.03  1.03    1.05  1.08  1.14 0.03
##              acc_b  0.36  0.43    0.46   0.51  0.50    0.54  0.58  0.68 0.05
##            acc_oob  0.33  0.41    0.45   0.49  0.48    0.52  0.56  0.66 0.05
##      sensitivity_b  0.06  0.41    0.54   0.61  0.63    0.76  0.84  0.99 0.15
##    sensitivity_oob  0.03  0.37    0.50   0.58  0.60    0.73  0.83  0.99 0.15
##      specificity_b  0.02  0.22    0.31   0.46  0.44    0.54  0.66  0.97 0.15
##    specificity_oob  0.02  0.20    0.30   0.44  0.42    0.53  0.64  0.95 0.15
##     cohens_kappa_b  0.01  0.03    0.05   0.06  0.06    0.08  0.10  0.15 0.02
##   cohens_kappa_oob -0.08 -0.02    0.01   0.02  0.02    0.04  0.07  0.12 0.03
##              ppv_b  0.32  0.34    0.36   0.37  0.37    0.38  0.40  0.50 0.02
##            ppv_oob  0.23  0.31    0.34   0.35  0.35    0.37  0.39  0.42 0.02
##              npv_b  0.64  0.67    0.69   0.70  0.70    0.72  0.74  0.89 0.02
##            npv_oob  0.58  0.64    0.66   0.68  0.68    0.69  0.72  0.83 0.03
##  NAs
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0

# ROC plot
plot(PER.NEU.IMI)

# Youden index plot
plot_metric(PER.NEU.IMI)

# LYM IMI cow

LYM.IMI <-  cutpointr(Pc.by.quarter.C3, log.Lym, IMI, 
                      method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(LYM.IMI)

## Method: maximize_metric 
## Predictor: log.Lym 
## Outcome: IMI 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC    n n_pos n_neg
##  0.5235 1594   544  1050
## 
##  optimal_cutpoint sum_sens_spec  acc sensitivity specificity    ppv    npv  tp
##            2.8904        1.0542 0.51      0.5809      0.4733 0.3636 0.6855 316
##   fn  fp  tn
##  228 553 497
## 
## Predictor summary: 
##     Data Min.       5%  1st Qu.   Median     Mean  3rd Qu.      95%     Max.
##  Overall    0 1.609438 2.419648 2.944439 3.001466 3.465736 4.488636 7.219642
##        0    0 1.609438 2.397895 2.890372 2.964868 3.433987 4.356709 7.195187
##        1    0 1.609438 2.484907 2.995732 3.072105 3.526361 4.850854 7.219642
##         SD NAs
##  0.9397576   0
##  0.8925654   0
##  1.0218769   0
## 
## Bootstrap summary: 
##           Variable  Min.    5% 1st Qu. Median Mean 3rd Qu.  95% Max.   SD NAs
##   optimal_cutpoint  1.95  2.83    2.89   3.00 3.31    3.56 4.49 6.58 0.64   0
##              AUC_b  0.47  0.50    0.51   0.52 0.52    0.53 0.55 0.57 0.02   0
##            AUC_oob  0.45  0.49    0.51   0.52 0.52    0.54 0.56 0.60 0.02   0
##    sum_sens_spec_b  1.01  1.03    1.05   1.06 1.06    1.08 1.10 1.14 0.02   0
##  sum_sens_spec_oob  0.92  0.97    1.00   1.03 1.02    1.04 1.07 1.14 0.03   0
##              acc_b  0.39  0.50    0.52   0.53 0.56    0.62 0.67 0.70 0.06   0
##            acc_oob  0.35  0.46    0.50   0.52 0.54    0.60 0.66 0.71 0.07   0
##      sensitivity_b  0.02  0.08    0.25   0.54 0.44    0.59 0.64 0.92 0.22   0
##    sensitivity_oob  0.02  0.06    0.21   0.51 0.41    0.57 0.62 0.88 0.21   0
##      specificity_b  0.11  0.43    0.48   0.53 0.63    0.81 0.97 1.00 0.21   0
##    specificity_oob  0.07  0.41    0.47   0.51 0.61    0.79 0.96 0.99 0.21   0
##     cohens_kappa_b  0.01  0.03    0.05   0.06 0.06    0.07 0.09 0.13 0.02   0
##   cohens_kappa_oob -0.08 -0.03    0.00   0.02 0.02    0.04 0.07 0.12 0.03   0
##              ppv_b  0.32  0.35    0.37   0.38 0.41    0.42 0.57 0.72 0.07   0
##            ppv_oob  0.23  0.31    0.34   0.36 0.37    0.38 0.50 0.69 0.06   0
##              npv_b  0.64  0.66    0.67   0.68 0.69    0.70 0.72 0.76 0.02   0
##            npv_oob  0.53  0.63    0.65   0.67 0.67    0.68 0.70 0.74 0.02   0

# ROC plot
plot(LYM.IMI)

# Youden index plot
plot_metric(LYM.IMI)

# %LYM IMI cow
PER.LYM.IMI <- cutpointr(Pc.by.quarter.C3, PERCENT02_N, IMI, 
                         method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(PER.LYM.IMI)

## Method: maximize_metric 
## Predictor: PERCENT02_N 
## Outcome: IMI 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC    n n_pos n_neg
##  0.5014 1594   544  1050
## 
##  optimal_cutpoint sum_sens_spec   acc sensitivity specificity    ppv    npv  tp
##              17.3        1.0312 0.468      0.6654      0.3657 0.3521 0.6784 362
##   fn  fp  tn
##  182 666 384
## 
## Predictor summary: 
##     Data Min.      5% 1st Qu. Median     Mean 3rd Qu.     95%  Max.       SD
##  Overall 2.70 10.3265 15.5700 19.305 19.67462 23.4400 29.6705 48.30 5.978316
##        0 2.70 10.7270 15.4925 19.230 19.69315 23.4775 29.7605 46.05 6.018691
##        1 4.53 10.0000 15.8575 19.420 19.63886 23.3350 29.4740 48.30 5.904945
##  NAs
##    0
##    0
##    0
## 
## Bootstrap summary: 
##           Variable  Min.    5% 1st Qu. Median  Mean 3rd Qu.   95%  Max.   SD
##   optimal_cutpoint  4.53 13.69   16.35  17.31 18.87   19.41 27.83 48.30 4.53
##              AUC_b  0.45  0.48    0.49   0.50  0.50    0.51  0.53  0.55 0.02
##            AUC_oob  0.43  0.47    0.49   0.50  0.50    0.51  0.53  0.56 0.02
##    sum_sens_spec_b  1.00  1.01    1.03   1.04  1.04    1.06  1.08  1.12 0.02
##  sum_sens_spec_oob  0.87  0.94    0.98   1.00  1.00    1.01  1.04  1.09 0.03
##              acc_b  0.31  0.41    0.46   0.48  0.50    0.53  0.65  0.68 0.07
##            acc_oob  0.32  0.39    0.44   0.46  0.48    0.50  0.63  0.68 0.07
##      sensitivity_b  0.00  0.10    0.52   0.67  0.58    0.73  0.85  1.00 0.23
##    sensitivity_oob  0.00  0.08    0.48   0.64  0.55    0.70  0.83  1.00 0.24
##      specificity_b  0.00  0.18    0.32   0.38  0.46    0.52  0.92  1.00 0.23
##    specificity_oob  0.00  0.16    0.30   0.36  0.44    0.50  0.91  1.00 0.23
##     cohens_kappa_b  0.00  0.01    0.03   0.04  0.04    0.05  0.07  0.10 0.02
##   cohens_kappa_oob -0.11 -0.05   -0.02   0.00 -0.01    0.01  0.04  0.07 0.03
##              ppv_b  0.31  0.34    0.35   0.36  0.37    0.38  0.42  1.00 0.04
##            ppv_oob  0.00  0.28    0.32   0.34  0.34    0.35  0.37  0.50 0.03
##              npv_b  0.63  0.66    0.67   0.69  0.69    0.70  0.72  1.00 0.03
##            npv_oob  0.33  0.61    0.64   0.66  0.66    0.67  0.70  1.00 0.03
##  NAs
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    4
##    0
##    5

# ROC plot
plot(PER.LYM.IMI)

# Youden index plot
plot_metric(PER.LYM.IMI)

# MAC IMI cow
MAC.IMI <- cutpointr(Pc.by.quarter.C3, log.Mac, IMI, 
                     method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 0

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(MAC.IMI)

## Method: maximize_metric 
## Predictor: log.Mac 
## Outcome: IMI 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC    n n_pos n_neg
##  0.4962 1594  1050   544
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv
##            2.7081        1.0215 0.5439      0.6152      0.4062 0.6667 0.3536
##   tp  fn  fp  tn
##  646 404 323 221
## 
## Predictor summary: 
##     Data Min.       5%  1st Qu.   Median     Mean  3rd Qu.      95%     Max.
##  Overall    0 1.609438 2.326413 2.890372 2.992118 3.496508 4.779472 8.986447
##        0    0 1.609438 2.397895 2.890372 2.971666 3.496508 4.626742 8.905309
##        1    0 1.386294 2.302585 2.890372 3.031595 3.526361 5.116021 8.986447
##        SD NAs
##  1.070470   0
##  1.004755   0
##  1.187056   0
## 
## Bootstrap summary: 
##           Variable  Min.    5% 1st Qu. Median  Mean 3rd Qu.  95% Max.   SD NAs
##   optimal_cutpoint  0.69  1.61    1.79   2.48   Inf    2.71 3.78  Inf  NaN   0
##              AUC_b  0.45  0.47    0.49   0.50  0.50    0.51 0.52 0.55 0.02   0
##            AUC_oob  0.43  0.46    0.48   0.50  0.50    0.51 0.53 0.56 0.02   0
##    sum_sens_spec_b  1.00  1.01    1.02   1.03  1.04    1.05 1.07 1.12 0.02   0
##  sum_sens_spec_oob  0.89  0.94    0.97   0.99  0.99    1.01 1.04 1.09 0.03   0
##              acc_b  0.33  0.41    0.55   0.58  0.57    0.63 0.66 0.69 0.07   0
##            acc_oob  0.30  0.38    0.52   0.55  0.55    0.62 0.66 0.69 0.08   0
##      sensitivity_b  0.00  0.18    0.61   0.71  0.68    0.91 0.97 1.00 0.23   0
##    sensitivity_oob  0.00  0.15    0.59   0.69  0.67    0.90 0.96 1.00 0.23   0
##      specificity_b  0.01  0.05    0.12   0.33  0.35    0.44 0.86 1.00 0.23   0
##    specificity_oob  0.00  0.04    0.10   0.31  0.32    0.41 0.82 1.00 0.22   0
##     cohens_kappa_b  0.00  0.01    0.02   0.03  0.04    0.05 0.07 0.12 0.02   0
##   cohens_kappa_oob -0.11 -0.05   -0.02  -0.01 -0.01    0.01 0.04 0.09 0.03   0
##              ppv_b  0.63  0.65    0.66   0.67  0.67    0.68 0.70 1.00 0.02   4
##            ppv_oob  0.00  0.60    0.64   0.65  0.65    0.67 0.69 0.71 0.03   4
##              npv_b  0.31  0.34    0.36   0.37  0.38    0.39 0.48 0.80 0.05   0
##            npv_oob  0.00  0.29    0.32   0.34  0.34    0.36 0.43 0.58 0.05   1

# ROC plot
plot(MAC.IMI)

## Warning: Removed 4 rows containing non-finite outside the scale range
## (`stat_density()`).

# Youden index plot
plot_metric(MAC.IMI)

# %MAC Major IMI cow
PER.MAC.IMI<- cutpointr(Pc.by.quarter.C3, PERCENT03_N, IMI, 
                        method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 0

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(PER.MAC.IMI)

## Method: maximize_metric 
## Predictor: PERCENT03_N 
## Outcome: IMI 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##    AUC    n n_pos n_neg
##  0.523 1594  1050   544
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv
##             26.45         1.058 0.4536      0.2914      0.7665 0.7067 0.3592
##   tp  fn  fp  tn
##  306 744 127 417
## 
## Predictor summary: 
##     Data Min.     5% 1st Qu. Median     Mean 3rd Qu.     95%  Max.       SD NAs
##  Overall 1.89 7.1530 13.0425 19.310 21.31286   27.27 42.5310 72.94 11.05134   0
##        0 2.00 7.1735 13.1600 19.600 21.66627   27.77 42.3275 72.94 11.28930   0
##        1 1.89 6.9980 12.7525 18.795 20.63074   26.06 42.6460 57.35 10.55365   0
## 
## Bootstrap summary: 
##           Variable  Min.    5% 1st Qu. Median  Mean 3rd Qu.   95%  Max.   SD
##   optimal_cutpoint  4.40 11.76   19.01  25.53 23.91   26.47 29.85 50.90 5.53
##              AUC_b  0.48  0.50    0.51   0.52  0.52    0.53  0.55  0.58 0.01
##            AUC_oob  0.46  0.49    0.51   0.52  0.52    0.54  0.56  0.60 0.02
##    sum_sens_spec_b  1.01  1.04    1.06   1.07  1.07    1.08  1.10  1.15 0.02
##  sum_sens_spec_oob  0.91  0.97    1.01   1.03  1.03    1.05  1.08  1.13 0.03
##              acc_b  0.34  0.42    0.45   0.47  0.49    0.53  0.62  0.68 0.06
##            acc_oob  0.33  0.41    0.43   0.45  0.47    0.50  0.59  0.68 0.05
##      sensitivity_b  0.02  0.22    0.28   0.32  0.39    0.54  0.82  0.99 0.18
##    sensitivity_oob  0.02  0.20    0.27   0.31  0.38    0.51  0.80  1.00 0.18
##      specificity_b  0.02  0.24    0.54   0.75  0.68    0.79  0.85  1.00 0.18
##    specificity_oob  0.01  0.20    0.50   0.73  0.65    0.78  0.84  0.99 0.19
##     cohens_kappa_b  0.01  0.03    0.05   0.06  0.06    0.07  0.09  0.13 0.02
##   cohens_kappa_oob -0.09 -0.03    0.00   0.02  0.02    0.04  0.06  0.10 0.03
##              ppv_b  0.64  0.67    0.69   0.71  0.71    0.73  0.75  0.91 0.03
##            ppv_oob  0.56  0.64    0.66   0.68  0.68    0.70  0.74  0.79 0.03
##              npv_b  0.32  0.34    0.36   0.37  0.37    0.38  0.40  0.66 0.03
##            npv_oob  0.11  0.31    0.34   0.35  0.35    0.36  0.38  0.67 0.02
##  NAs
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0

# ROC plot
plot(PER.MAC.IMI)

# Youden index plot
plot_metric(PER.MAC.IMI)

# Summary of results when using C5 and IMI as reference

print(c("Optimal SCC:", exp(SCC.IMI$optimal_cutpoint), SCC.IMI$sensitivity, SCC.IMI$specificity, SCC.IMI$ppv, SCC.IMI$npv, SCC.IMI$AUC))

##                                                 sensitivity         specificity 
##      "Optimal SCC:"                "43" "0.912045889101338" "0.115808823529412" 
##                                                             
## "0.664808362369338" "0.406451612903226" "0.494115362445169"

print(c("Optimal n NEU:", exp(NEU.IMI$optimal_cutpoint), NEU.IMI$sensitivity, NEU.IMI$specificity, NEU.IMI$ppv, NEU.IMI$npv, NEU.IMI$AUC))

##                                                 sensitivity         specificity 
##    "Optimal n NEU:"                "38" "0.751838235294118"               "0.3" 
##                                                             
## "0.357517482517482"               "0.7" "0.527754726890756"

print(c("Optimal %NEU:", PER.NEU.IMI$optimal_cutpoint, PER.NEU.IMI$sensitivity, PER.NEU.IMI$specificity, PER.NEU.IMI$ppv, PER.NEU.IMI$npv, PER.NEU.IMI$AUC))

##                                                 sensitivity         specificity 
##     "Optimal %NEU:"             "59.65" "0.542279411764706" "0.515238095238095" 
##                                                             
## "0.366915422885572" "0.684810126582278" "0.526458333333333"

print(c("Optimal n LYM:", exp(LYM.IMI$optimal_cutpoint), LYM.IMI$sensitivity, LYM.IMI$specificity, LYM.IMI$ppv, LYM.IMI$npv, LYM.IMI$AUC))

##                                                 sensitivity         specificity 
##    "Optimal n LYM:"                "18" "0.580882352941177" "0.473333333333333" 
##                                                             
## "0.363636363636364"  "0.68551724137931" "0.523480392156863"

print(c("Optimal %LYM:", PER.LYM.IMI$optimal_cutpoint, PER.LYM.IMI$sensitivity, PER.LYM.IMI$specificity, PER.LYM.IMI$ppv, PER.LYM.IMI$npv, PER.LYM.IMI$AUC))

##                                                 sensitivity         specificity 
##     "Optimal %LYM:"              "17.3" "0.665441176470588" "0.365714285714286" 
##                                                             
## "0.352140077821012" "0.678445229681979" "0.501424194677871"

print(c("Optimal n MAC:", exp(MAC.IMI$optimal_cutpoint), MAC.IMI$sensitivity, MAC.IMI$specificity, MAC.IMI$ppv, MAC.IMI$npv, MAC.IMI$AUC))

##                                                 sensitivity         specificity 
##    "Optimal n MAC:"                "15" "0.615238095238095"           "0.40625" 
##                                                             
## "0.666666666666667"            "0.3536" "0.496233368347339"

print(c("Optimal %MAC:", PER.MAC.IMI$optimal_cutpoint, PER.MAC.IMI$sensitivity, PER.MAC.IMI$specificity, PER.MAC.IMI$ppv, PER.MAC.IMI$npv, PER.MAC.IMI$AUC))

##                                                 sensitivity         specificity 
##     "Optimal %MAC:"             "26.45" "0.291428571428571" "0.766544117647059" 
##                                                             
## "0.706697459584296" "0.359173126614987" "0.523040966386555"

# SCC.IMI
SCC.IMI_optimal_cutpoint <- boot_ci(SCC.IMI, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
SCC.IMI_lower_ci <- subset(SCC.IMI_optimal_cutpoint, SCC.IMI_optimal_cutpoint$quantile == 0.025)$values
exp(SCC.IMI_lower_ci)

## [1] 27

SCC.IMI_upper_ci <- subset(SCC.IMI_optimal_cutpoint, SCC.IMI_optimal_cutpoint$quantile == 0.975)$values
exp(SCC.IMI_upper_ci)

## [1] 139

boot_ci(SCC.IMI, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.395
## 2    0.975  0.988

boot_ci(SCC.IMI, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025 0.0314
## 2    0.975 0.651

boot_ci(SCC.IMI, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.645
## 2    0.975  0.700

boot_ci(SCC.IMI, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.345
## 2    0.975  0.629

boot_ci(SCC.IMI, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.464
## 2    0.975  0.525

# NEU.IMI
NEU.IMI_optimal_cutpoint <- boot_ci(NEU.IMI, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
NEU.IMI_lower_ci <- subset(NEU.IMI_optimal_cutpoint, NEU.IMI_optimal_cutpoint$quantile == 0.025)$values
exp(NEU.IMI_lower_ci)

## [1] 36

NEU.IMI_upper_ci <- subset(NEU.IMI_optimal_cutpoint, NEU.IMI_optimal_cutpoint$quantile == 0.975)$values
exp(NEU.IMI_upper_ci)

## [1] 269

boot_ci(NEU.IMI, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025 0.0844
## 2    0.975 0.797

boot_ci(NEU.IMI, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.274
## 2    0.975  0.958

boot_ci(NEU.IMI, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.340
## 2    0.975  0.536

boot_ci(NEU.IMI, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.656
## 2    0.975  0.741

boot_ci(NEU.IMI, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.497
## 2    0.975  0.558

# PER.NEU.IMI
PER.NEU.IMI_optimal_cutpoint <- boot_ci(PER.NEU.IMI, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
PER.NEU.IMI_optimal_cutpoint

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025   47.4
## 2    0.975   67.1

PER.NEU.IMI_lower_ci <- subset(PER.NEU.IMI_optimal_cutpoint, PER.NEU.IMI_optimal_cutpoint$quantile == 0.025)$values
PER.NEU.IMI_upper_ci <- subset(PER.NEU.IMI_optimal_cutpoint, PER.NEU.IMI_optimal_cutpoint$quantile == 0.975)$values

boot_ci(PER.NEU.IMI, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.317
## 2    0.975  0.865

boot_ci(PER.NEU.IMI, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.192
## 2    0.975  0.743

boot_ci(PER.NEU.IMI, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.339
## 2    0.975  0.411

boot_ci(PER.NEU.IMI, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.664
## 2    0.975  0.752

boot_ci(PER.NEU.IMI, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.497
## 2    0.975  0.555

# LYM.IMI
LYM.IMI_optimal_cutpoint <- boot_ci(LYM.IMI, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
LYM.IMI_lower_ci <- subset(LYM.IMI_optimal_cutpoint, LYM.IMI_optimal_cutpoint$quantile == 0.025)$values
exp(LYM.IMI_lower_ci)

## [1] 13

LYM.IMI_upper_ci <- subset(LYM.IMI_optimal_cutpoint, LYM.IMI_optimal_cutpoint$quantile == 0.975)$values
exp(LYM.IMI_upper_ci)

## [1] 105

boot_ci(LYM.IMI, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025 0.0662
## 2    0.975 0.768

boot_ci(LYM.IMI, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.291
## 2    0.975  0.973

boot_ci(LYM.IMI, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.345
## 2    0.975  0.6

boot_ci(LYM.IMI, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.653
## 2    0.975  0.723

boot_ci(LYM.IMI, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.493
## 2    0.975  0.554

# PER.LYM.IMI
PER.LYM.IMI_optimal_cutpoint <- boot_ci(PER.LYM.IMI, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
PER.LYM.IMI_optimal_cutpoint

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025   13.5
## 2    0.975   28.0

PER.LYM.IMI_lower_ci <- subset(PER.LYM.IMI_optimal_cutpoint, PER.LYM.IMI_optimal_cutpoint$quantile == 0.025)$values
PER.LYM.IMI_upper_ci <- subset(PER.LYM.IMI_optimal_cutpoint, PER.LYM.IMI_optimal_cutpoint$quantile == 0.975)$values

boot_ci(PER.LYM.IMI, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025 0.0922
## 2    0.975 0.885

boot_ci(PER.LYM.IMI, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.151
## 2    0.975  0.930

boot_ci(PER.LYM.IMI, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.330
## 2    0.975  0.445

boot_ci(PER.LYM.IMI, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.651
## 2    0.975  0.739

boot_ci(PER.LYM.IMI, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.471
## 2    0.975  0.532

# MAC.IMI
MAC.IMI_optimal_cutpoint <- boot_ci(MAC.IMI, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
MAC.IMI_lower_ci <- subset(MAC.IMI_optimal_cutpoint, MAC.IMI_optimal_cutpoint$quantile == 0.025)$values
exp(MAC.IMI_lower_ci)

## [1] 5

MAC.IMI_upper_ci <- subset(MAC.IMI_optimal_cutpoint, MAC.IMI_optimal_cutpoint$quantile == 0.975)$values
exp(MAC.IMI_upper_ci)

## [1] 57

boot_ci(MAC.IMI, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.140
## 2    0.975  0.970

boot_ci(MAC.IMI, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025 0.0436
## 2    0.975 0.884

boot_ci(MAC.IMI, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.645
## 2    0.975  0.714

boot_ci(MAC.IMI, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.334
## 2    0.975  0.5

boot_ci(MAC.IMI, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.467
## 2    0.975  0.526

# PER.MAC.IMI
PER.MAC.IMI_optimal_cutpoint <- boot_ci(PER.MAC.IMI, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
PER.MAC.IMI_optimal_cutpoint

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025   10.7
## 2    0.975   33.3

PER.MAC.IMI_lower_ci <- subset(PER.MAC.IMI_optimal_cutpoint, PER.MAC.IMI_optimal_cutpoint$quantile == 0.025)$values
PER.MAC.IMI_upper_ci <- subset(PER.MAC.IMI_optimal_cutpoint, PER.MAC.IMI_optimal_cutpoint$quantile == 0.975)$values

boot_ci(PER.MAC.IMI, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.166
## 2    0.975  0.856

boot_ci(PER.MAC.IMI, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.201
## 2    0.975  0.883

boot_ci(PER.MAC.IMI, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.664
## 2    0.975  0.758

boot_ci(PER.MAC.IMI, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.338
## 2    0.975  0.423

boot_ci(PER.MAC.IMI, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.494
## 2    0.975  0.551

# Prevalence

# True prevalence

table(Pc.by.quarter.C3$IMI)

## 
##    0    1 
## 1050  545

prop.table(table(Pc.by.quarter.C3$IMI))

## 
##         0         1 
## 0.6583072 0.3416928

#Apparent prevalence

# ---- SCC ----

# optimal cut-off

optimal_scc_C3_IMI_quarter <- SCC.IMI$optimal_cutpoint

table(Pc.by.quarter.C3$test_result_n > exp(optimal_scc_C3_IMI_quarter))

## 
## FALSE  TRUE 
##   179  1501

prop.table(table(Pc.by.quarter.C3$test_result_n > exp(optimal_scc_C3_IMI_quarter)))

## 
##     FALSE      TRUE 
## 0.1065476 0.8934524

# confidence intervals

table(Pc.by.quarter.C3$test_result_n > exp(SCC.IMI_lower_ci))

## 
## FALSE  TRUE 
##    48  1632

prop.table(table(Pc.by.quarter.C3$test_result_n > exp(SCC.IMI_lower_ci)))

## 
##      FALSE       TRUE 
## 0.02857143 0.97142857

table(Pc.by.quarter.C3$test_result_n > exp(SCC.IMI_upper_ci))

## 
## FALSE  TRUE 
##  1033   647

prop.table(table(Pc.by.quarter.C3$test_result_n >exp(SCC.IMI_upper_ci)))

## 
##    FALSE     TRUE 
## 0.614881 0.385119

# ---- Neutrophils ----

# optimal cut-off

optimal_neu_C3_IMI_quarter <- NEU.IMI$optimal_cutpoint

table(Pc.by.quarter.C3$TOTAL01_N > exp(optimal_neu_C3_IMI_quarter))

## 
## FALSE  TRUE 
##   477  1207

prop.table(table(Pc.by.quarter.C3$TOTAL01_N > exp(optimal_neu_C3_IMI_quarter)))

## 
##     FALSE      TRUE 
## 0.2832542 0.7167458

# confidence intervals

table(Pc.by.quarter.C3$TOTAL01_N > exp(NEU.IMI_lower_ci))

## 
## FALSE  TRUE 
##   452  1232

prop.table(table(Pc.by.quarter.C3$TOTAL01_N > exp(NEU.IMI_lower_ci)))

## 
##     FALSE      TRUE 
## 0.2684086 0.7315914

table(Pc.by.quarter.C3$TOTAL01_N > exp(NEU.IMI_upper_ci))

## 
## FALSE  TRUE 
##  1595    89

prop.table(table(Pc.by.quarter.C3$TOTAL01_N >exp(NEU.IMI_upper_ci)))

## 
##      FALSE       TRUE 
## 0.94714964 0.05285036

# ---- Percent neu ----

# Optimal cut-off

optimal_per_neu_C3_IMI_quarter <- PER.NEU.IMI$optimal_cutpoint

table(Pc.by.quarter.C3$PERCENT01_N > optimal_per_neu_C3_IMI_quarter)

## 
## FALSE  TRUE 
##   841   843

prop.table(table(Pc.by.quarter.C3$PERCENT01_N > optimal_per_neu_C3_IMI_quarter))

## 
##     FALSE      TRUE 
## 0.4994062 0.5005938

# confidence intervals

table(Pc.by.quarter.C3$PERCENT01_N > PER.NEU.IMI_lower_ci)

## 
## FALSE  TRUE 
##   282  1402

prop.table(table(Pc.by.quarter.C3$PERCENT01_N > PER.NEU.IMI_lower_ci))

## 
##     FALSE      TRUE 
## 0.1674584 0.8325416

table(Pc.by.quarter.C3$PERCENT01_N > PER.NEU.IMI_upper_ci)

## 
## FALSE  TRUE 
##  1248   436

prop.table(table(Pc.by.quarter.C3$PERCENT01_N >PER.NEU.IMI_upper_ci))

## 
##     FALSE      TRUE 
## 0.7410926 0.2589074

# ---- Lymphocytes ----

# optimal cut-off

optimal_lym_C3_IMI_quarter <-LYM.IMI$optimal_cutpoint

table(Pc.by.quarter.C3$TOTAL02_N > exp(optimal_lym_C3_IMI_quarter))

## 
## FALSE  TRUE 
##   764   920

prop.table(table(Pc.by.quarter.C3$TOTAL02_N > exp(optimal_lym_C3_IMI_quarter)))

## 
##     FALSE      TRUE 
## 0.4536817 0.5463183

# confidence intervals

table(Pc.by.quarter.C3$TOTAL02_N > exp(LYM.IMI_lower_ci))

## 
## FALSE  TRUE 
##   542  1142

prop.table(table(Pc.by.quarter.C3$TOTAL02_N > exp(LYM.IMI_lower_ci)))

## 
##     FALSE      TRUE 
## 0.3218527 0.6781473

table(Pc.by.quarter.C3$TOTAL02_N > exp(LYM.IMI_upper_ci))

## 
## FALSE  TRUE 
##  1619    65

prop.table(table(Pc.by.quarter.C3$TOTAL02_N >exp(LYM.IMI_upper_ci)))

## 
##      FALSE       TRUE 
## 0.96140143 0.03859857

# ---- Percent lym ----

# optimal cut-off

optimal_per_lym_C3_IMI_quarter <-PER.LYM.IMI$optimal_cutpoint

table(Pc.by.quarter.C3$PERCENT02_N > optimal_per_lym_C3_IMI_quarter)

## 
## FALSE  TRUE 
##   609  1075

prop.table(table(Pc.by.quarter.C3$PERCENT02_N > optimal_per_lym_C3_IMI_quarter))

## 
##    FALSE     TRUE 
## 0.361639 0.638361

# confidence intervals

table(Pc.by.quarter.C3$PERCENT02_N > PER.LYM.IMI_lower_ci)

## 
## FALSE  TRUE 
##   249  1435

prop.table(table(Pc.by.quarter.C3$PERCENT02_N > PER.LYM.IMI_lower_ci))

## 
##     FALSE      TRUE 
## 0.1478622 0.8521378

table(Pc.by.quarter.C3$PERCENT02_N > PER.LYM.IMI_upper_ci)

## 
## FALSE  TRUE 
##  1545   139

prop.table(table(Pc.by.quarter.C3$PERCENT02_N >PER.LYM.IMI_upper_ci))

## 
##      FALSE       TRUE 
## 0.91745843 0.08254157

# ---- Macrophages ----

# optimal cut-off

optimal_mac_C3_IMI_quarter <- MAC.IMI$optimal_cutpoint

table(Pc.by.quarter.C3$TOTAL03_N > exp(optimal_mac_C3_IMI_quarter))

## 
## FALSE  TRUE 
##   701   983

prop.table(table(Pc.by.quarter.C3$TOTAL03_N > exp(optimal_mac_C3_IMI_quarter)))

## 
##     FALSE      TRUE 
## 0.4162708 0.5837292

# confidence intervals

table(Pc.by.quarter.C3$TOTAL03_N > exp(MAC.IMI_lower_ci))

## 
## FALSE  TRUE 
##    75  1609

prop.table(table(Pc.by.quarter.C3$TOTAL03_N > exp(MAC.IMI_lower_ci)))

## 
##      FALSE       TRUE 
## 0.04453682 0.95546318

table(Pc.by.quarter.C3$TOTAL03_N > exp(MAC.IMI_upper_ci))

## 
## FALSE  TRUE 
##  1463   221

prop.table(table(Pc.by.quarter.C3$TOTAL03_N >exp(MAC.IMI_upper_ci)))

## 
##     FALSE      TRUE 
## 0.8687648 0.1312352

# Percent MAC

# optimal cut-off

optimal_per_mac_C3_IMI_quarter <- PER.MAC.IMI$optimal_cutpoint

table(Pc.by.quarter.C3$PERCENT03_N > optimal_per_mac_C3_IMI_quarter)

## 
## FALSE  TRUE 
##  1223   461

prop.table(table(Pc.by.quarter.C3$PERCENT03_N > optimal_per_mac_C3_IMI_quarter))

## 
##    FALSE     TRUE 
## 0.726247 0.273753

# confidence intervals

table(Pc.by.quarter.C3$PERCENT03_N > PER.MAC.IMI_lower_ci)

## 
## FALSE  TRUE 
##   268  1416

prop.table(table(Pc.by.quarter.C3$PERCENT03_N > PER.MAC.IMI_lower_ci))

## 
##     FALSE      TRUE 
## 0.1591449 0.8408551

table(Pc.by.quarter.C3$PERCENT03_N > PER.MAC.IMI_upper_ci)

## 
## FALSE  TRUE 
##  1446   238

prop.table(table(Pc.by.quarter.C3$PERCENT03_N >PER.MAC.IMI_upper_ci))

## 
##     FALSE      TRUE 
## 0.8586698 0.1413302

Quarter high SCC

# Set a seed for reproducibility

set.seed(123)

# Qscout subclinical mastitis reference

# n NEU SCM quarter
NEU.delaval.SCM <- cutpointr(Pc.by.quarter.C3, log.Neu, delaval.SCM, 
                                 method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(NEU.delaval.SCM)

## Method: maximize_metric 
## Predictor: log.Neu 
## Outcome: delaval.SCM 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC    n n_pos n_neg
##  0.8559 1679   365  1314
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity   ppv    npv  tp
##            4.4427        1.5665 0.8017      0.7507      0.8158 0.531 0.9218 274
##  fn  fp   tn
##  91 242 1072
## 
## Predictor summary: 
##     Data     Min.       5%  1st Qu.   Median     Mean  3rd Qu.      95%
##  Overall 1.609438 2.833213 3.555348 4.043051 4.134314 4.595120 5.650025
##        0 1.609438 2.772589 3.401197 3.871201 3.860068 4.290459 5.022112
##        1 2.833213 3.806662 4.442651 4.927254 5.121600 5.480639 7.497096
##      Max.        SD NAs
##  9.429396 0.9643956   0
##  7.668094 0.6986908   0
##  9.429396 1.1303368   0
## 
## Bootstrap summary: 
##           Variable Min.   5% 1st Qu. Median Mean 3rd Qu.  95% Max.   SD NAs
##   optimal_cutpoint 4.22 4.38    4.44   4.45 4.46    4.50 4.52 4.71 0.05   0
##              AUC_b 0.82 0.84    0.85   0.86 0.86    0.86 0.87 0.90 0.01   0
##            AUC_oob 0.81 0.83    0.85   0.86 0.86    0.87 0.88 0.90 0.01   0
##    sum_sens_spec_b 1.49 1.53    1.55   1.57 1.57    1.59 1.61 1.68 0.02   0
##  sum_sens_spec_oob 1.41 1.50    1.53   1.55 1.55    1.58 1.61 1.67 0.03   0
##              acc_b 0.73 0.78    0.80   0.81 0.81    0.82 0.83 0.85 0.01   0
##            acc_oob 0.69 0.77    0.79   0.80 0.80    0.81 0.83 0.85 0.02   0
##      sensitivity_b 0.61 0.71    0.73   0.75 0.75    0.77 0.80 0.86 0.03   0
##    sensitivity_oob 0.53 0.68    0.71   0.74 0.74    0.76 0.80 0.85 0.04   0
##      specificity_b 0.71 0.78    0.81   0.82 0.82    0.83 0.85 0.90 0.02   0
##    specificity_oob 0.66 0.78    0.80   0.82 0.82    0.83 0.85 0.91 0.02   0
##     cohens_kappa_b 0.38 0.46    0.48   0.50 0.50    0.52 0.54 0.58 0.03   0
##   cohens_kappa_oob 0.34 0.43    0.46   0.49 0.49    0.51 0.54 0.60 0.03   0
##              ppv_b 0.42 0.49    0.52   0.54 0.54    0.56 0.58 0.65 0.03   0
##            ppv_oob 0.40 0.47    0.51   0.53 0.53    0.55 0.59 0.67 0.03   0
##              npv_b 0.88 0.91    0.92   0.92 0.92    0.93 0.94 0.96 0.01   0
##            npv_oob 0.86 0.90    0.91   0.92 0.92    0.93 0.94 0.95 0.01   0

# ROC plot
plot(NEU.delaval.SCM)

# Youden index plot
plot_metric(NEU.delaval.SCM)

# %NEU SCM quarter
PER.NEU.delaval.SCM <- cutpointr(Pc.by.quarter.C3, PERCENT01_N, delaval.SCM, 
                                 method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 0

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(PER.NEU.delaval.SCM)

## Method: maximize_metric 
## Predictor: PERCENT01_N 
## Outcome: delaval.SCM 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##    AUC    n n_pos n_neg
##  0.582 1679  1314   365
## 
##  optimal_cutpoint sum_sens_spec   acc sensitivity specificity    ppv    npv
##             46.43        1.1572 0.757      0.8942       0.263 0.8137 0.4085
##    tp  fn  fp tn
##  1175 139 269 96
## 
## Predictor summary: 
##     Data  Min.     5% 1st Qu. Median     Mean 3rd Qu.    95%  Max.       SD NAs
##  Overall 20.16 38.910  51.365 59.650 58.93719  67.410 76.400 87.31 11.46029   0
##        0 21.28 41.670  52.315 60.195 59.79792  67.815 76.400 87.31 10.81338   0
##        1 20.16 33.364  46.100 57.140 55.83858  65.210 75.878 84.56 13.09094   0
## 
## Bootstrap summary: 
##           Variable  Min.    5% 1st Qu. Median  Mean 3rd Qu.   95%  Max.   SD
##   optimal_cutpoint 41.38 43.20   46.43  46.43 47.98   47.62 61.10 65.30 4.73
##              AUC_b  0.52  0.55    0.57   0.58  0.58    0.59  0.61  0.64 0.02
##            AUC_oob  0.51  0.55    0.57   0.58  0.58    0.60  0.62  0.66 0.02
##    sum_sens_spec_b  1.08  1.12    1.15   1.16  1.16    1.18  1.20  1.25 0.02
##  sum_sens_spec_oob  1.00  1.07    1.11   1.14  1.13    1.16  1.19  1.27 0.04
##              acc_b  0.45  0.53    0.74   0.76  0.73    0.77  0.78  0.81 0.07
##            acc_oob  0.41  0.49    0.73   0.75  0.72    0.76  0.78  0.81 0.08
##      sensitivity_b  0.37  0.49    0.87   0.89  0.85    0.90  0.94  0.96 0.13
##    sensitivity_oob  0.32  0.46    0.87   0.89  0.84    0.90  0.93  0.96 0.14
##      specificity_b  0.16  0.21    0.25   0.27  0.31    0.30  0.67  0.82 0.13
##    specificity_oob  0.12  0.19    0.23   0.26  0.29    0.29  0.62  0.75 0.12
##     cohens_kappa_b  0.06  0.10    0.16   0.18  0.18    0.20  0.22  0.28 0.04
##   cohens_kappa_oob  0.00  0.05    0.13   0.16  0.15    0.18  0.22  0.31 0.05
##              ppv_b  0.78  0.80    0.81   0.82  0.82    0.82  0.84  0.88 0.01
##            ppv_oob  0.76  0.79    0.80   0.81  0.81    0.82  0.83  0.85 0.01
##              npv_b  0.22  0.26    0.38   0.41  0.40    0.44  0.48  0.58 0.06
##            npv_oob  0.18  0.24    0.34   0.39  0.38    0.42  0.47  0.60 0.07
##  NAs
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0

# ROC plot
plot(PER.NEU.delaval.SCM)

# Youden index plot
plot_metric(PER.NEU.delaval.SCM)

# LYM SCM quarter
LYM.delaval.SCM <- cutpointr(Pc.by.quarter.C3, log.Lym, delaval.SCM, method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(LYM.delaval.SCM)

## Method: maximize_metric 
## Predictor: log.Lym 
## Outcome: delaval.SCM 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC    n n_pos n_neg
##  0.8959 1679   365  1314
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv
##            3.3673        1.6472 0.8338      0.8055      0.8417 0.5857 0.9397
##   tp fn  fp   tn
##  294 71 208 1106
## 
## Predictor summary: 
##     Data     Min.       5%  1st Qu.   Median     Mean  3rd Qu.      95%
##  Overall 0.000000 1.609438 2.484907 2.944439 3.007395 3.496508 4.488636
##        0 0.000000 1.609438 2.302585 2.772589 2.715713 3.178054 3.737670
##        1 1.609438 2.784714 3.465736 3.891820 4.057452 4.406719 6.473399
##      Max.        SD NAs
##  7.219642 0.9427204   0
##  6.626718 0.6824364   0
##  7.219642 1.0021081   0
## 
## Bootstrap summary: 
##           Variable Min.   5% 1st Qu. Median Mean 3rd Qu.  95% Max.   SD NAs
##   optimal_cutpoint 3.18 3.26    3.37   3.37 3.37    3.37 3.47 3.58 0.06   0
##              AUC_b 0.86 0.88    0.89   0.90 0.90    0.90 0.91 0.92 0.01   0
##            AUC_oob 0.84 0.88    0.89   0.90 0.90    0.91 0.92 0.93 0.01   0
##    sum_sens_spec_b 1.56 1.61    1.63   1.65 1.65    1.66 1.68 1.74 0.02   0
##  sum_sens_spec_oob 1.53 1.58    1.62   1.64 1.64    1.66 1.69 1.75 0.03   0
##              acc_b 0.77 0.80    0.83   0.84 0.83    0.84 0.86 0.89 0.02   0
##            acc_oob 0.76 0.79    0.82   0.83 0.83    0.84 0.85 0.89 0.02   0
##      sensitivity_b 0.68 0.76    0.79   0.81 0.81    0.83 0.86 0.91 0.03   0
##    sensitivity_oob 0.64 0.73    0.78   0.80 0.80    0.82 0.86 0.92 0.04   0
##      specificity_b 0.76 0.79    0.83   0.84 0.84    0.85 0.88 0.93 0.03   0
##    specificity_oob 0.74 0.78    0.83   0.84 0.84    0.85 0.88 0.93 0.03   0
##     cohens_kappa_b 0.46 0.52    0.55   0.57 0.57    0.59 0.62 0.67 0.03   0
##   cohens_kappa_oob 0.42 0.50    0.54   0.56 0.56    0.58 0.61 0.68 0.03   0
##              ppv_b 0.48 0.53    0.57   0.59 0.59    0.61 0.65 0.76 0.04   0
##            ppv_oob 0.43 0.51    0.56   0.58 0.58    0.61 0.64 0.72 0.04   0
##              npv_b 0.91 0.93    0.93   0.94 0.94    0.95 0.95 0.97 0.01   0
##            npv_oob 0.89 0.92    0.93   0.94 0.94    0.95 0.96 0.97 0.01   0

# ROC plot
plot(LYM.delaval.SCM)

# Youden index plot
plot_metric(LYM.delaval.SCM)

# %LYM SCM quarter
PER.LYM.delaval.SCM <- cutpointr(Pc.by.quarter.C3, PERCENT02_N, delaval.SCM, 
                                 method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(PER.LYM.delaval.SCM)

## Method: maximize_metric 
## Predictor: PERCENT02_N 
## Outcome: delaval.SCM 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC    n n_pos n_neg
##  0.5001 1679   365  1314
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv
##             15.96        1.0385 0.3853      0.7562      0.2823 0.2264 0.8065
##   tp fn  fp  tn
##  276 89 943 371
## 
## Predictor summary: 
##     Data Min.      5% 1st Qu. Median     Mean 3rd Qu.    95% Max.       SD NAs
##  Overall 2.70 10.3290 15.5700  19.25 19.64905  23.385 29.703 48.3 5.974163   0
##        0 2.70 10.2795 15.4575  19.23 19.69201  23.455 30.230 48.3 6.126172   0
##        1 3.97 10.8440 16.0000  19.35 19.49438  23.030 28.876 35.7 5.396589   0
## 
## Bootstrap summary: 
##           Variable  Min.    5% 1st Qu. Median  Mean 3rd Qu.   95%  Max.   SD
##   optimal_cutpoint  3.97 12.70   13.85  15.96 16.10   18.81 19.06 28.82 2.51
##              AUC_b  0.45  0.47    0.49   0.50  0.50    0.51  0.53  0.56 0.02
##            AUC_oob  0.42  0.46    0.48   0.50  0.50    0.51  0.54  0.57 0.02
##    sum_sens_spec_b  1.00  1.02    1.04   1.05  1.05    1.07  1.09  1.13 0.02
##  sum_sens_spec_oob  0.86  0.95    0.99   1.01  1.01    1.03  1.06  1.09 0.03
##              acc_b  0.21  0.29    0.33   0.39  0.40    0.49  0.52  0.76 0.08
##            acc_oob  0.19  0.28    0.33   0.37  0.39    0.46  0.50  0.74 0.08
##      sensitivity_b  0.07  0.55    0.61   0.78  0.75    0.86  0.91  1.00 0.14
##    sensitivity_oob  0.01  0.48    0.56   0.74  0.71    0.84  0.90  1.00 0.15
##      specificity_b  0.00  0.12    0.18   0.28  0.31    0.46  0.50  0.94 0.14
##    specificity_oob  0.00  0.11    0.18   0.27  0.30    0.44  0.49  0.93 0.14
##     cohens_kappa_b  0.00  0.01    0.02   0.03  0.03    0.04  0.06  0.09 0.01
##   cohens_kappa_oob -0.13 -0.03   -0.01   0.00  0.00    0.02  0.03  0.06 0.02
##              ppv_b  0.19  0.21    0.22   0.23  0.23    0.24  0.25  0.31 0.01
##            ppv_oob  0.06  0.19    0.21   0.22  0.22    0.23  0.24  0.27 0.02
##              npv_b  0.77  0.79    0.80   0.82  0.82    0.83  0.86  1.00 0.02
##            npv_oob  0.00  0.75    0.77   0.79  0.79    0.81  0.84  1.00 0.04
##  NAs
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    2

# ROC plot
plot(PER.LYM.delaval.SCM)

# Youden index plot
plot_metric(PER.LYM.delaval.SCM)

# MAC SCM quarter
MAC.delaval.SCM <- cutpointr(Pc.by.quarter.C3, log.Mac, delaval.SCM, method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(MAC.delaval.SCM)

## Method: maximize_metric 
## Predictor: log.Mac 
## Outcome: delaval.SCM 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##    AUC    n n_pos n_neg
##  0.878 1679   365  1314
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv
##            3.4012        1.5951 0.8148      0.7671       0.828 0.5534 0.9275
##   tp fn  fp   tn
##  280 85 226 1088
## 
## Predictor summary: 
##     Data     Min.       5%  1st Qu.   Median     Mean  3rd Qu.      95%
##  Overall 0.000000 1.609438 2.397895 2.944439 3.004812 3.526361 4.773195
##        0 0.000000 1.386294 2.197225 2.708050 2.683366 3.178054 3.850148
##        1 1.609438 2.639057 3.401197 4.025352 4.162018 4.644391 6.726675
##      Max.        SD NAs
##  8.986447 1.0721298   0
##  7.136483 0.7725831   0
##  8.986447 1.1958438   0
## 
## Bootstrap summary: 
##           Variable Min.   5% 1st Qu. Median Mean 3rd Qu.  95% Max.   SD NAs
##   optimal_cutpoint 3.00 3.22    3.40   3.40 3.40    3.47 3.53 3.66 0.09   0
##              AUC_b 0.84 0.86    0.87   0.88 0.88    0.88 0.89 0.91 0.01   0
##            AUC_oob 0.83 0.86    0.87   0.88 0.88    0.89 0.90 0.92 0.01   0
##    sum_sens_spec_b 1.52 1.56    1.58   1.60 1.60    1.62 1.64 1.68 0.02   0
##  sum_sens_spec_oob 1.44 1.52    1.56   1.58 1.58    1.60 1.64 1.69 0.03   0
##              acc_b 0.70 0.77    0.81   0.82 0.82    0.83 0.85 0.87 0.02   0
##            acc_oob 0.68 0.76    0.80   0.82 0.81    0.83 0.84 0.87 0.02   0
##      sensitivity_b 0.64 0.71    0.74   0.76 0.77    0.79 0.84 0.91 0.04   0
##    sensitivity_oob 0.54 0.67    0.72   0.75 0.75    0.78 0.83 0.91 0.05   0
##      specificity_b 0.64 0.75    0.82   0.84 0.83    0.86 0.88 0.92 0.04   0
##    specificity_oob 0.63 0.74    0.81   0.83 0.83    0.85 0.88 0.92 0.04   0
##     cohens_kappa_b 0.37 0.46    0.51   0.53 0.53    0.55 0.58 0.64 0.04   0
##   cohens_kappa_oob 0.34 0.44    0.49   0.52 0.51    0.54 0.57 0.64 0.04   0
##              ppv_b 0.40 0.48    0.54   0.57 0.56    0.59 0.63 0.72 0.04   0
##            ppv_oob 0.38 0.47    0.53   0.55 0.55    0.58 0.63 0.70 0.05   0
##              npv_b 0.90 0.91    0.92   0.93 0.93    0.93 0.95 0.96 0.01   0
##            npv_oob 0.87 0.90    0.92   0.92 0.92    0.93 0.94 0.97 0.01   0

# ROC plot
plot(MAC.delaval.SCM)

# Youden index plot
plot_metric(MAC.delaval.SCM)

# %MAC SCM quarter
PER.MAC.delaval.SCM<- cutpointr(Pc.by.quarter.C3, PERCENT03_N, delaval.SCM, method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(PER.MAC.delaval.SCM)

## Method: maximize_metric 
## Predictor: PERCENT03_N 
## Outcome: delaval.SCM 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC    n n_pos n_neg
##  0.5758 1679   365  1314
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv
##             30.39        1.1658 0.7296      0.3233      0.8425 0.3631 0.8176
##   tp  fn  fp   tn
##  118 247 207 1107
## 
## Predictor summary: 
##     Data Min.     5% 1st Qu. Median     Mean 3rd Qu.     95%  Max.       SD NAs
##  Overall 1.89 7.1870  13.110  19.39 21.41483  27.385 43.0030 72.94 11.08192   0
##        0 1.89 7.3465  12.995  18.98 20.51173  26.335 38.4035 67.91 10.05564   0
##        1 1.94 7.1360  13.600  21.95 24.66600  34.300 49.7020 72.94 13.70595   0
## 
## Bootstrap summary: 
##           Variable  Min.    5% 1st Qu. Median  Mean 3rd Qu.   95%  Max.   SD
##   optimal_cutpoint 18.97 23.40   30.39  32.12 31.35   33.11 34.30 42.51 2.89
##              AUC_b  0.52  0.55    0.56   0.58  0.58    0.59  0.60  0.64 0.02
##            AUC_oob  0.48  0.54    0.56   0.58  0.58    0.59  0.61  0.67 0.02
##    sum_sens_spec_b  1.11  1.13    1.16   1.17  1.17    1.19  1.21  1.27 0.02
##  sum_sens_spec_oob  1.01  1.08    1.12   1.15  1.15    1.17  1.20  1.27 0.04
##              acc_b  0.53  0.64    0.73   0.75  0.74    0.76  0.77  0.80 0.04
##            acc_oob  0.49  0.61    0.72   0.74  0.73    0.75  0.77  0.81 0.04
##      sensitivity_b  0.15  0.25    0.28   0.31  0.32    0.34  0.49  0.66 0.07
##    sensitivity_oob  0.14  0.22    0.26   0.29  0.30    0.32  0.43  0.53 0.06
##      specificity_b  0.50  0.68    0.85   0.87  0.86    0.89  0.91  0.97 0.06
##    specificity_oob  0.49  0.66    0.84   0.87  0.85    0.89  0.91  0.98 0.07
##     cohens_kappa_b  0.09  0.14    0.17   0.19  0.19    0.20  0.23  0.29 0.03
##   cohens_kappa_oob  0.00  0.07    0.13   0.16  0.16    0.19  0.22  0.32 0.04
##              ppv_b  0.26  0.30    0.36   0.39  0.39    0.42  0.46  0.60 0.05
##            ppv_oob  0.20  0.26    0.33   0.37  0.37    0.40  0.45  0.65 0.05
##              npv_b  0.78  0.80    0.81   0.82  0.82    0.83  0.84  0.86 0.01
##            npv_oob  0.77  0.79    0.80   0.81  0.81    0.82  0.84  0.87 0.01
##  NAs
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0

# ROC plot
plot(PER.MAC.delaval.SCM)

# Youden index plot
plot_metric(PER.MAC.delaval.SCM)

# Summary of results when using SCM as reference

print(c("Optimal n NEU:", exp(NEU.delaval.SCM$optimal_cutpoint), NEU.delaval.SCM$sensitivity, NEU.delaval.SCM$specificity, NEU.delaval.SCM$ppv, NEU.delaval.SCM$npv, NEU.delaval.SCM$AUC))

##                                                 sensitivity         specificity 
##    "Optimal n NEU:"                "85" "0.750684931506849" "0.815829528158295" 
##                                                             
## "0.531007751937985" "0.921754084264832" "0.855941285627906"

print(c("Optimal %NEU:", PER.NEU.delaval.SCM$optimal_cutpoint, PER.NEU.delaval.SCM$sensitivity, PER.NEU.delaval.SCM$specificity, PER.NEU.delaval.SCM$ppv, PER.NEU.delaval.SCM$npv, PER.NEU.delaval.SCM$AUC))

##                                                 sensitivity         specificity 
##     "Optimal %NEU:"             "46.43" "0.894216133942161" "0.263013698630137" 
##                                                             
## "0.813711911357341" "0.408510638297872" "0.582028106169596"

print(c("Optimal n LYM:", exp(LYM.delaval.SCM$optimal_cutpoint), LYM.delaval.SCM$sensitivity, LYM.delaval.SCM$specificity, LYM.delaval.SCM$ppv, LYM.delaval.SCM$npv, LYM.delaval.SCM$AUC))

##                                                 sensitivity         specificity 
##    "Optimal n LYM:"                "29" "0.805479452054795" "0.841704718417047" 
##                                                             
## "0.585657370517928" "0.939677145284622" "0.895938366589521"

print(c("Optimal %LYM:", PER.LYM.delaval.SCM$optimal_cutpoint, PER.LYM.delaval.SCM$sensitivity, PER.LYM.delaval.SCM$specificity, PER.LYM.delaval.SCM$ppv, PER.LYM.delaval.SCM$npv, PER.LYM.delaval.SCM$AUC))

##                                                 sensitivity         specificity 
##     "Optimal %LYM:"             "15.96" "0.756164383561644"  "0.28234398782344" 
##                                                             
## "0.226415094339623" "0.806521739130435"  "0.50011050645316"

print(c("Optimal n MAC:", exp(MAC.delaval.SCM$optimal_cutpoint), MAC.delaval.SCM$sensitivity, MAC.delaval.SCM$specificity, MAC.delaval.SCM$ppv, MAC.delaval.SCM$npv, MAC.delaval.SCM$AUC))

##                                                 sensitivity         specificity 
##    "Optimal n MAC:"                "30" "0.767123287671233" "0.828006088280061" 
##                                                             
## "0.553359683794466" "0.927536231884058"  "0.87804049123246"

print(c("Optimal %MAC:", PER.MAC.delaval.SCM$optimal_cutpoint, PER.MAC.delaval.SCM$sensitivity, PER.MAC.delaval.SCM$specificity, PER.MAC.delaval.SCM$ppv, PER.MAC.delaval.SCM$npv, PER.MAC.delaval.SCM$AUC))

##                                                 sensitivity         specificity 
##     "Optimal %MAC:"             "30.39" "0.323287671232877" "0.842465753424658" 
##                                                             
## "0.363076923076923" "0.817577548005908"  "0.57584391484748"

# NEU.delaval.SCM
NEU.delaval.SCM_optimal_cutpoint <- boot_ci(NEU.delaval.SCM, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
NEU.delaval.SCM_lower_ci <- subset(NEU.delaval.SCM_optimal_cutpoint, NEU.delaval.SCM_optimal_cutpoint$quantile == 0.025)$values
exp(NEU.delaval.SCM_lower_ci)

## [1] 80

NEU.delaval.SCM_upper_ci <- subset(NEU.delaval.SCM_optimal_cutpoint, NEU.delaval.SCM_optimal_cutpoint$quantile == 0.975)$values
exp(NEU.delaval.SCM_upper_ci)

## [1] 93

boot_ci(NEU.delaval.SCM, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.698
## 2    0.975  0.806

boot_ci(NEU.delaval.SCM, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.776
## 2    0.975  0.855

boot_ci(NEU.delaval.SCM, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.480
## 2    0.975  0.594

boot_ci(NEU.delaval.SCM, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.906
## 2    0.975  0.939

boot_ci(NEU.delaval.SCM, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.835
## 2    0.975  0.876

# PER.NEU.delaval.SCM
PER.NEU.delaval.SCM_optimal_cutpoint <- boot_ci(PER.NEU.delaval.SCM, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
PER.NEU.delaval.SCM_optimal_cutpoint

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025   42.8
## 2    0.975   61.2

PER.NEU.delaval.SCM_lower_ci <- subset(PER.NEU.delaval.SCM_optimal_cutpoint, PER.NEU.delaval.SCM_optimal_cutpoint$quantile == 0.025)$values
PER.NEU.delaval.SCM_upper_ci <- subset(PER.NEU.delaval.SCM_optimal_cutpoint, PER.NEU.delaval.SCM_optimal_cutpoint$quantile == 0.975)$values

boot_ci(PER.NEU.delaval.SCM, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.471
## 2    0.975  0.942

boot_ci(PER.NEU.delaval.SCM, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.196
## 2    0.975  0.693

boot_ci(PER.NEU.delaval.SCM, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.795
## 2    0.975  0.850

boot_ci(PER.NEU.delaval.SCM, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.256
## 2    0.975  0.507

boot_ci(PER.NEU.delaval.SCM, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.546
## 2    0.975  0.614

# LYM.delaval.SCM
LYM.delaval.SCM_optimal_cutpoint <- boot_ci(LYM.delaval.SCM, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
LYM.delaval.SCM_lower_ci <- subset(LYM.delaval.SCM_optimal_cutpoint, LYM.delaval.SCM_optimal_cutpoint$quantile == 0.025)$values
exp(LYM.delaval.SCM_lower_ci)

## [1] 26

LYM.delaval.SCM_upper_ci <- subset(LYM.delaval.SCM_optimal_cutpoint, LYM.delaval.SCM_optimal_cutpoint$quantile == 0.975)$values
exp(LYM.delaval.SCM_upper_ci)

## [1] 33

boot_ci(LYM.delaval.SCM, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.746
## 2    0.975  0.867

boot_ci(LYM.delaval.SCM, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.783
## 2    0.975  0.892

boot_ci(LYM.delaval.SCM, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.514
## 2    0.975  0.667

boot_ci(LYM.delaval.SCM, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.923
## 2    0.975  0.958

boot_ci(LYM.delaval.SCM, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.877
## 2    0.975  0.914

# PER.LYM.delaval.SCM
PER.LYM.delaval.SCM_optimal_cutpoint <- boot_ci(PER.LYM.delaval.SCM, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
PER.LYM.delaval.SCM_optimal_cutpoint

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025   10.6
## 2    0.975   20.9

PER.LYM.delaval.SCM_lower_ci <- subset(PER.LYM.delaval.SCM_optimal_cutpoint, PER.LYM.delaval.SCM_optimal_cutpoint$quantile == 0.025)$values
PER.LYM.delaval.SCM_upper_ci <- subset(PER.LYM.delaval.SCM_optimal_cutpoint, PER.LYM.delaval.SCM_optimal_cutpoint$quantile == 0.975)$values

boot_ci(PER.LYM.delaval.SCM, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.463
## 2    0.975  0.962

boot_ci(PER.LYM.delaval.SCM, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025 0.0613
## 2    0.975 0.604

boot_ci(PER.LYM.delaval.SCM, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.209
## 2    0.975  0.256

boot_ci(PER.LYM.delaval.SCM, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.785
## 2    0.975  0.870

boot_ci(PER.LYM.delaval.SCM, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.470
## 2    0.975  0.532

# MAC.delaval.SCM
MAC.delaval.SCM_optimal_cutpoint <- boot_ci(MAC.delaval.SCM, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
MAC.delaval.SCM_lower_ci <- subset(MAC.delaval.SCM_optimal_cutpoint, MAC.delaval.SCM_optimal_cutpoint$quantile == 0.025)$values
exp(MAC.delaval.SCM_lower_ci)

## [1] 24

MAC.delaval.SCM_upper_ci <- subset(MAC.delaval.SCM_optimal_cutpoint, MAC.delaval.SCM_optimal_cutpoint$quantile == 0.975)$values
exp(MAC.delaval.SCM_upper_ci)

## [1] 34

boot_ci(MAC.delaval.SCM, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.702
## 2    0.975  0.855

boot_ci(MAC.delaval.SCM, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.735
## 2    0.975  0.885

boot_ci(MAC.delaval.SCM, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.465
## 2    0.975  0.642

boot_ci(MAC.delaval.SCM, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.910
## 2    0.975  0.950

boot_ci(MAC.delaval.SCM, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.857
## 2    0.975  0.898

# PER.MAC.delaval.SCM
PER.MAC.delaval.SCM_optimal_cutpoint <- boot_ci(PER.MAC.delaval.SCM, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
PER.MAC.delaval.SCM_optimal_cutpoint

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025   22.9
## 2    0.975   34.7

PER.MAC.delaval.SCM_lower_ci <- subset(PER.MAC.delaval.SCM_optimal_cutpoint, PER.MAC.delaval.SCM_optimal_cutpoint$quantile == 0.025)$values
PER.MAC.delaval.SCM_upper_ci <- subset(PER.MAC.delaval.SCM_optimal_cutpoint, PER.MAC.delaval.SCM_optimal_cutpoint$quantile == 0.975)$values

boot_ci(PER.MAC.delaval.SCM, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.235
## 2    0.975  0.514

boot_ci(PER.MAC.delaval.SCM, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.662
## 2    0.975  0.918

boot_ci(PER.MAC.delaval.SCM, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.290
## 2    0.975  0.476

boot_ci(PER.MAC.delaval.SCM, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.799
## 2    0.975  0.840

boot_ci(PER.MAC.delaval.SCM, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.542
## 2    0.975  0.609

# Prevalence

# True prevalence

table(Pc.by.quarter.C3$delaval.SCM)

## 
##    0    1 
## 1314  366

prop.table(table(Pc.by.quarter.C3$delaval.SCM))

## 
##         0         1 
## 0.7821429 0.2178571

#Apparent prevalence

# ---- Neutrophils ----

# optimal cut-off

optimal_neu_C3_delaval.SCM_quarter <- NEU.delaval.SCM$optimal_cutpoint

table(Pc.by.quarter.C3$TOTAL01_N > exp(optimal_neu_C3_delaval.SCM_quarter))

## 
## FALSE  TRUE 
##  1173   511

prop.table(table(Pc.by.quarter.C3$TOTAL01_N > exp(optimal_neu_C3_delaval.SCM_quarter)))

## 
##     FALSE      TRUE 
## 0.6965558 0.3034442

# confidence intervals

table(Pc.by.quarter.C3$TOTAL01_N > exp(NEU.delaval.SCM_lower_ci))

## 
## FALSE  TRUE 
##  1113   571

prop.table(table(Pc.by.quarter.C3$TOTAL01_N > exp(NEU.delaval.SCM_lower_ci)))

## 
##     FALSE      TRUE 
## 0.6609264 0.3390736

table(Pc.by.quarter.C3$TOTAL01_N > exp(NEU.delaval.SCM_upper_ci))

## 
## FALSE  TRUE 
##  1227   457

prop.table(table(Pc.by.quarter.C3$TOTAL01_N >exp(NEU.delaval.SCM_upper_ci)))

## 
##     FALSE      TRUE 
## 0.7286223 0.2713777

# ---- Percent neu ----

# Optimal cut-off

optimal_per_neu_C3_delaval.SCM_quarter <- PER.NEU.delaval.SCM$optimal_cutpoint

table(Pc.by.quarter.C3$PERCENT01_N > optimal_per_neu_C3_delaval.SCM_quarter)

## 
## FALSE  TRUE 
##   239  1445

prop.table(table(Pc.by.quarter.C3$PERCENT01_N > optimal_per_neu_C3_delaval.SCM_quarter))

## 
##    FALSE     TRUE 
## 0.141924 0.858076

# confidence intervals

table(Pc.by.quarter.C3$PERCENT01_N > PER.NEU.delaval.SCM_lower_ci)

## 
## FALSE  TRUE 
##   151  1533

prop.table(table(Pc.by.quarter.C3$PERCENT01_N > PER.NEU.delaval.SCM_lower_ci))

## 
##      FALSE       TRUE 
## 0.08966746 0.91033254

table(Pc.by.quarter.C3$PERCENT01_N > PER.NEU.delaval.SCM_upper_ci)

## 
## FALSE  TRUE 
##   941   743

prop.table(table(Pc.by.quarter.C3$PERCENT01_N >PER.NEU.delaval.SCM_upper_ci))

## 
##     FALSE      TRUE 
## 0.5587886 0.4412114

# ---- Lymphocytes ----

# optimal cut-off

optimal_lym_C3_delaval.SCM_quarter <-LYM.delaval.SCM$optimal_cutpoint

table(Pc.by.quarter.C3$TOTAL02_N > exp(optimal_lym_C3_delaval.SCM_quarter))

## 
## FALSE  TRUE 
##  1204   480

prop.table(table(Pc.by.quarter.C3$TOTAL02_N > exp(optimal_lym_C3_delaval.SCM_quarter)))

## 
##     FALSE      TRUE 
## 0.7149644 0.2850356

# confidence intervals

table(Pc.by.quarter.C3$TOTAL02_N > exp(LYM.delaval.SCM_lower_ci))

## 
## FALSE  TRUE 
##  1136   548

prop.table(table(Pc.by.quarter.C3$TOTAL02_N > exp(LYM.delaval.SCM_lower_ci)))

## 
##     FALSE      TRUE 
## 0.6745843 0.3254157

table(Pc.by.quarter.C3$TOTAL02_N > exp(LYM.delaval.SCM_upper_ci))

## 
## FALSE  TRUE 
##  1278   406

prop.table(table(Pc.by.quarter.C3$TOTAL02_N >exp(LYM.delaval.SCM_upper_ci)))

## 
##     FALSE      TRUE 
## 0.7589074 0.2410926

# ---- Percent lym ----

# optimal cut-off

optimal_per_lym_C3_delaval.SCM_quarter <-PER.LYM.delaval.SCM$optimal_cutpoint

table(Pc.by.quarter.C3$PERCENT02_N > optimal_per_lym_C3_delaval.SCM_quarter)

## 
## FALSE  TRUE 
##   463  1221

prop.table(table(Pc.by.quarter.C3$PERCENT02_N > optimal_per_lym_C3_delaval.SCM_quarter))

## 
##     FALSE      TRUE 
## 0.2749406 0.7250594

# confidence intervals

table(Pc.by.quarter.C3$PERCENT02_N > PER.LYM.delaval.SCM_lower_ci)

## 
## FALSE  TRUE 
##    95  1589

prop.table(table(Pc.by.quarter.C3$PERCENT02_N > PER.LYM.delaval.SCM_lower_ci))

## 
##     FALSE      TRUE 
## 0.0564133 0.9435867

table(Pc.by.quarter.C3$PERCENT02_N > PER.LYM.delaval.SCM_upper_ci)

## 
## FALSE  TRUE 
##  1018   666

prop.table(table(Pc.by.quarter.C3$PERCENT02_N >PER.LYM.delaval.SCM_upper_ci))

## 
##     FALSE      TRUE 
## 0.6045131 0.3954869

# ---- Macrophages ----

# optimal cut-off

optimal_mac_C3_delaval.SCM_quarter <- MAC.delaval.SCM$optimal_cutpoint

table(Pc.by.quarter.C3$TOTAL03_N > exp(optimal_mac_C3_delaval.SCM_quarter))

## 
## FALSE  TRUE 
##  1209   475

prop.table(table(Pc.by.quarter.C3$TOTAL03_N > exp(optimal_mac_C3_delaval.SCM_quarter)))

## 
##     FALSE      TRUE 
## 0.7179335 0.2820665

# confidence intervals

table(Pc.by.quarter.C3$TOTAL03_N > exp(MAC.delaval.SCM_lower_ci))

## 
## FALSE  TRUE 
##  1057   627

prop.table(table(Pc.by.quarter.C3$TOTAL03_N > exp(MAC.delaval.SCM_lower_ci)))

## 
##     FALSE      TRUE 
## 0.6276722 0.3723278

table(Pc.by.quarter.C3$TOTAL03_N > exp(MAC.delaval.SCM_upper_ci))

## 
## FALSE  TRUE 
##  1280   404

prop.table(table(Pc.by.quarter.C3$TOTAL03_N >exp(MAC.delaval.SCM_upper_ci)))

## 
##    FALSE     TRUE 
## 0.760095 0.239905

# Percent MAC

# optimal cut-off

optimal_per_mac_C3_delaval.SCM_quarter <- PER.MAC.delaval.SCM$optimal_cutpoint

table(Pc.by.quarter.C3$PERCENT03_N > optimal_per_mac_C3_delaval.SCM_quarter)

## 
## FALSE  TRUE 
##  1360   324

prop.table(table(Pc.by.quarter.C3$PERCENT03_N > optimal_per_mac_C3_delaval.SCM_quarter))

## 
##    FALSE     TRUE 
## 0.807601 0.192399

# confidence intervals

table(Pc.by.quarter.C3$PERCENT03_N > PER.MAC.delaval.SCM_lower_ci)

## 
## FALSE  TRUE 
##  1050   634

prop.table(table(Pc.by.quarter.C3$PERCENT03_N > PER.MAC.delaval.SCM_lower_ci))

## 
##     FALSE      TRUE 
## 0.6235154 0.3764846

table(Pc.by.quarter.C3$PERCENT03_N > PER.MAC.delaval.SCM_upper_ci)

## 
## FALSE  TRUE 
##  1484   200

prop.table(table(Pc.by.quarter.C3$PERCENT03_N > PER.MAC.delaval.SCM_upper_ci))

## 
##     FALSE      TRUE 
## 0.8812352 0.1187648

Postpartum C5

Cow IMI

# Set a seed for reproducibility

set.seed(123)

# SCC IMI cow

SCC.IMI <- cutpointr(Pc.by.cow.C5, log.SCC.CO.COW, IMI.cow, 
                     method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(SCC.IMI)

## Method: maximize_metric 
## Predictor: log.SCC.CO.COW 
## Outcome: IMI.cow 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC   n n_pos n_neg
##  0.5467 417   309   108
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv tp
##             4.804        1.1547 0.4317      0.2751      0.8796 0.8673 0.2978 85
##   fn fp tn
##  224 13 95
## 
## Predictor summary: 
##     Data     Min.       5%  1st Qu.   Median     Mean  3rd Qu.      95%
##  Overall 1.386294 3.081738 3.737670 4.234107 4.454499 4.762174 7.297509
##        0 2.484907 3.232603 3.707399 4.151009 4.280960 4.579813 6.001713
##        1 1.386294 2.995732 3.761200 4.248495 4.515153 4.867534 7.514084
##      Max.        SD NAs
##  8.572249 1.1737622   0
##  8.203030 0.8937674   0
##  8.572249 1.2526175   0
## 
## Bootstrap summary: 
##           Variable  Min.    5% 1st Qu. Median Mean 3rd Qu.  95% Max.   SD NAs
##   optimal_cutpoint  3.74  4.04    4.71   4.80 4.74    4.85 4.99 6.31 0.26   0
##              AUC_b  0.45  0.50    0.53   0.55 0.55    0.57 0.60 0.65 0.03   0
##            AUC_oob  0.38  0.48    0.52   0.55 0.55    0.57 0.61 0.68 0.04   0
##    sum_sens_spec_b  1.05  1.10    1.14   1.17 1.17    1.19 1.23 1.32 0.04   0
##  sum_sens_spec_oob  0.82  0.97    1.07   1.12 1.11    1.15 1.21 1.34 0.07   0
##              acc_b  0.29  0.39    0.42   0.44 0.46    0.47 0.61 0.68 0.06   0
##            acc_oob  0.31  0.37    0.41   0.43 0.44    0.46 0.54 0.64 0.05   0
##      sensitivity_b  0.11  0.22    0.26   0.29 0.32    0.33 0.64 0.79 0.11   0
##    sensitivity_oob  0.09  0.20    0.25   0.28 0.31    0.32 0.59 0.75 0.11   0
##      specificity_b  0.26  0.54    0.84   0.88 0.85    0.91 0.94 0.99 0.11   0
##    specificity_oob  0.21  0.40    0.79   0.85 0.80    0.89 0.94 0.98 0.15   0
##     cohens_kappa_b  0.03  0.06    0.09   0.11 0.11    0.12 0.16 0.26 0.03   0
##   cohens_kappa_oob -0.15 -0.03    0.04   0.07 0.07    0.10 0.13 0.23 0.05   0
##              ppv_b  0.73  0.80    0.84   0.87 0.87    0.89 0.92 0.97 0.04   0
##            ppv_oob  0.63  0.72    0.79   0.83 0.83    0.87 0.92 0.97 0.06   0
##              npv_b  0.22  0.26    0.29   0.30 0.30    0.32 0.35 0.40 0.03   0
##            npv_oob  0.13  0.22    0.26   0.29 0.29    0.31 0.35 0.41 0.04   0

# ROC plot
plot(SCC.IMI)

# Youden index plot
plot_metric(SCC.IMI)

# NEU IMI cow
NEU.IMI <- cutpointr(Pc.by.cow.C5, log.TOTAL.NEU.CO.COW, IMI.cow, 
                     method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(NEU.IMI)

## Method: maximize_metric 
## Predictor: log.TOTAL.NEU.CO.COW 
## Outcome: IMI.cow 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC   n n_pos n_neg
##  0.5369 419   311   108
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity   ppv    npv tp
##            4.2195        1.1457 0.4153      0.2476      0.8981 0.875 0.2931 77
##   fn fp tn
##  234 11 97
## 
## Predictor summary: 
##     Data     Min.       5%  1st Qu.   Median     Mean  3rd Qu.      95%
##  Overall 1.386294 2.302585 2.890372 3.401197 3.680466 4.068990 6.305914
##        0 1.945910 2.335944 2.944439 3.258097 3.528510 3.789808 5.274794
##        1 1.386294 2.249905 2.890372 3.433987 3.733235 4.197174 6.455302
##      Max.       SD NAs
##  8.948976 1.267673   0
##  8.350902 1.070034   0
##  8.948976 1.326957   0
## 
## Bootstrap summary: 
##           Variable  Min.    5% 1st Qu. Median Mean 3rd Qu.  95% Max.   SD NAs
##   optimal_cutpoint  3.00  3.30    3.87   4.11 4.02    4.22 4.25 4.36 0.28   0
##              AUC_b  0.44  0.49    0.52   0.54 0.54    0.56 0.59 0.65 0.03   0
##            AUC_oob  0.41  0.47    0.51   0.54 0.54    0.56 0.60 0.66 0.04   0
##    sum_sens_spec_b  1.03  1.10    1.13   1.16 1.16    1.19 1.23 1.29 0.04   0
##  sum_sens_spec_oob  0.80  0.97    1.06   1.11 1.10    1.15 1.20 1.29 0.07   0
##              acc_b  0.34  0.39    0.42   0.44 0.45    0.47 0.58 0.67 0.05   0
##            acc_oob  0.30  0.36    0.40   0.42 0.43    0.45 0.53 0.61 0.05   0
##      sensitivity_b  0.18  0.22    0.25   0.28 0.32    0.34 0.59 0.74 0.10   0
##    sensitivity_oob  0.13  0.20    0.24   0.27 0.30    0.32 0.56 0.70 0.10   0
##      specificity_b  0.34  0.57    0.83   0.88 0.84    0.91 0.94 0.98 0.10   0
##    specificity_oob  0.18  0.46    0.76   0.84 0.80    0.89 0.94 1.00 0.14   0
##     cohens_kappa_b  0.02  0.06    0.08   0.10 0.10    0.12 0.15 0.21 0.03   0
##   cohens_kappa_oob -0.15 -0.02    0.03   0.06 0.06    0.09 0.12 0.18 0.04   0
##              ppv_b  0.74  0.79    0.84   0.87 0.86    0.89 0.93 0.98 0.04   0
##            ppv_oob  0.64  0.72    0.78   0.83 0.82    0.87 0.92 1.00 0.06   0
##              npv_b  0.21  0.26    0.28   0.30 0.30    0.32 0.35 0.44 0.03   0
##            npv_oob  0.14  0.22    0.26   0.28 0.28    0.31 0.34 0.40 0.04   0

# ROC plot
plot(NEU.IMI)

# Youden index plot
plot_metric(NEU.IMI)

# %NEU IMI cow
PER.NEU.IMI <- cutpointr(Pc.by.cow.C5, PERCENT.NEU.CO.COW, IMI.cow, 
                         method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(PER.NEU.IMI)

## Method: maximize_metric 
## Predictor: PERCENT.NEU.CO.COW 
## Outcome: IMI.cow 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC   n n_pos n_neg
##  0.5037 419   311   108
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv
##             54.94        1.0687 0.5465      0.5595      0.5093 0.7665 0.2865
##   tp  fn fp tn
##  174 137 53 55
## 
## Predictor summary: 
##     Data  Min.    5% 1st Qu. Median     Mean 3rd Qu.    95%  Max.       SD NAs
##  Overall 18.75 37.08  49.465  56.45 56.18427  64.360 73.096 86.73 11.04959   0
##        0 27.33 37.67  50.000  54.68 56.14639  65.805 72.730 81.25 10.97584   0
##        1 18.75 37.30  49.325  57.07 56.19743  64.185 73.160 86.73 11.09267   0
## 
## Bootstrap summary: 
##           Variable  Min.    5% 1st Qu. Median Mean 3rd Qu.   95% Max.   SD NAs
##   optimal_cutpoint 27.61 43.48   54.76  55.81  Inf   57.83 68.38  Inf  NaN   0
##              AUC_b  0.39  0.45    0.48   0.50 0.50    0.53  0.56 0.61 0.03   0
##            AUC_oob  0.36  0.44    0.48   0.50 0.50    0.53  0.58 0.66 0.04   0
##    sum_sens_spec_b  1.00  1.02    1.06   1.09 1.09    1.12  1.17 1.25 0.05   0
##  sum_sens_spec_oob  0.77  0.89    0.96   1.00 1.00    1.04  1.11 1.22 0.06   0
##              acc_b  0.21  0.33    0.51   0.55 0.53    0.58  0.72 0.79 0.11   0
##            acc_oob  0.22  0.30    0.46   0.51 0.50    0.55  0.67 0.80 0.11   0
##      sensitivity_b  0.00  0.14    0.47   0.55 0.52    0.59  0.89 1.00 0.22   0
##    sensitivity_oob  0.00  0.11    0.43   0.52 0.49    0.57  0.87 1.00 0.22   0
##      specificity_b  0.01  0.17    0.51   0.56 0.57    0.64  0.94 1.00 0.21   0
##    specificity_oob  0.00  0.07    0.42   0.50 0.51    0.58  0.90 1.00 0.23   0
##     cohens_kappa_b  0.00  0.02    0.04   0.07 0.07    0.10  0.14 0.21 0.04   0
##   cohens_kappa_oob -0.19 -0.09   -0.03   0.00 0.00    0.03  0.08 0.18 0.05   0
##              ppv_b  0.69  0.74    0.76   0.78 0.79    0.81  0.87 1.00 0.05   1
##            ppv_oob  0.33  0.66    0.71   0.74 0.74    0.77  0.81 1.00 0.06  16
##              npv_b  0.19  0.25    0.27   0.30 0.31    0.32  0.39 1.00 0.08   0
##            npv_oob  0.00  0.16    0.23   0.26 0.25    0.28  0.33 1.00 0.07   5

# ROC plot
plot(PER.NEU.IMI)

## Warning: Removed 1 row containing non-finite outside the scale range
## (`stat_density()`).

# Youden index plot
plot_metric(PER.NEU.IMI)

# LYM IMI cow

LYM.IMI <-  cutpointr(Pc.by.cow.C5, log.TOTAL.LYM.CO.COW, IMI.cow, 
                      method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(LYM.IMI)

## Method: maximize_metric 
## Predictor: log.TOTAL.LYM.CO.COW 
## Outcome: IMI.cow 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC   n n_pos n_neg
##  0.5315 419   311   108
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity  ppv    npv tp
##            3.0445        1.1594 0.4344      0.2797      0.8796 0.87 0.2978 87
##   fn fp tn
##  224 13 95
## 
## Predictor summary: 
##     Data Min.       5% 1st Qu.   Median     Mean  3rd Qu.      95%     Max.
##  Overall    0 1.098612 1.94591 2.484907 2.640265 2.944439 5.162445 7.123673
##        0    0 1.386294 1.94591 2.397895 2.527277 2.833213 4.703667 6.562444
##        1    0 1.098612 1.94591 2.484907 2.679503 3.238486 5.176006 7.123673
##        SD NAs
##  1.180885   0
##  1.034037   0
##  1.226891   0
## 
## Bootstrap summary: 
##           Variable  Min.   5% 1st Qu. Median Mean 3rd Qu.  95% Max.   SD NAs
##   optimal_cutpoint  2.08 2.94    3.00   3.04 3.03    3.14 3.14 3.43 0.11   0
##              AUC_b  0.41 0.48    0.51   0.53 0.53    0.55 0.58 0.63 0.03   0
##            AUC_oob  0.40 0.46    0.50   0.53 0.53    0.56 0.59 0.66 0.04   0
##    sum_sens_spec_b  1.04 1.10    1.14   1.17 1.17    1.19 1.23 1.28 0.04   0
##  sum_sens_spec_oob  0.80 1.02    1.10   1.14 1.13    1.17 1.23 1.30 0.06   0
##              acc_b  0.37 0.40    0.42   0.44 0.44    0.46 0.49 0.66 0.03   0
##            acc_oob  0.32 0.38    0.41   0.43 0.43    0.45 0.48 0.57 0.03   0
##      sensitivity_b  0.18 0.24    0.27   0.29 0.29    0.31 0.35 0.77 0.05   0
##    sensitivity_oob  0.16 0.22    0.26   0.28 0.28    0.31 0.35 0.68 0.05   0
##      specificity_b  0.37 0.81    0.86   0.88 0.88    0.90 0.93 0.97 0.05   0
##    specificity_oob  0.16 0.72    0.82   0.86 0.85    0.89 0.94 1.00 0.08   0
##     cohens_kappa_b  0.02 0.06    0.09   0.10 0.10    0.12 0.15 0.19 0.03   0
##   cohens_kappa_oob -0.16 0.01    0.06   0.08 0.08    0.11 0.14 0.20 0.04   0
##              ppv_b  0.76 0.81    0.85   0.87 0.87    0.90 0.92 0.97 0.03   0
##            ppv_oob  0.64 0.76    0.82   0.85 0.85    0.89 0.93 1.00 0.05   0
##              npv_b  0.22 0.26    0.28   0.30 0.30    0.32 0.34 0.38 0.03   0
##            npv_oob  0.16 0.23    0.27   0.29 0.29    0.31 0.34 0.43 0.03   0

# ROC plot
plot(LYM.IMI)

# Youden index plot
plot_metric(LYM.IMI)

# %LYM IMI cow
PER.LYM.IMI <- cutpointr(Pc.by.cow.C5, PERCENT.LYM.CO.COW, IMI.cow, 
                         method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 0

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(PER.LYM.IMI)

## Method: maximize_metric 
## Predictor: PERCENT.LYM.CO.COW 
## Outcome: IMI.cow 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC   n n_pos n_neg
##  0.5362 419   108   311
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv tp
##             21.82        1.0924 0.5776      0.4815      0.6109 0.3006 0.7724 52
##  fn  fp  tn
##  56 121 190
## 
## Predictor summary: 
##     Data Min.      5% 1st Qu. Median     Mean 3rd Qu.     95%  Max.       SD
##  Overall  0.0 10.4470 15.7150  20.00 20.80084  25.250 34.2940 44.79 7.256602
##        0  6.9 10.3785 16.6175  21.29 21.41593  26.165 34.5185 38.60 6.990463
##        1  0.0 10.4700 15.3000  19.63 20.58723  24.415 34.1050 44.79 7.345701
##  NAs
##    0
##    0
##    0
## 
## Bootstrap summary: 
##           Variable  Min.    5% 1st Qu. Median  Mean 3rd Qu.   95%  Max.   SD
##   optimal_cutpoint  6.90 15.29   16.11  19.92 19.92   21.82 28.36 36.17 4.03
##              AUC_b  0.42  0.48    0.51   0.54  0.53    0.56  0.59  0.64 0.03
##            AUC_oob  0.39  0.47    0.51   0.54  0.54    0.57  0.61  0.67 0.04
##    sum_sens_spec_b  1.01  1.05    1.09   1.12  1.12    1.15  1.20  1.27 0.04
##  sum_sens_spec_oob  0.80  0.92    0.98   1.03  1.03    1.07  1.13  1.25 0.07
##              acc_b  0.21  0.38    0.44   0.55  0.53    0.60  0.69  0.77 0.10
##            acc_oob  0.19  0.35    0.42   0.51  0.50    0.57  0.65  0.79 0.09
##      sensitivity_b  0.05  0.25    0.50   0.60  0.61    0.80  0.88  1.00 0.19
##    sensitivity_oob  0.00  0.16    0.42   0.53  0.54    0.72  0.85  1.00 0.21
##      specificity_b  0.01  0.23    0.31   0.53  0.51    0.62  0.85  0.98 0.19
##    specificity_oob  0.01  0.21    0.29   0.51  0.48    0.61  0.82  0.98 0.19
##     cohens_kappa_b  0.00  0.04    0.06   0.09  0.09    0.12  0.16  0.25 0.04
##   cohens_kappa_oob -0.17 -0.07   -0.01   0.02  0.02    0.05  0.10  0.20 0.05
##              ppv_b  0.20  0.25    0.28   0.31  0.31    0.33  0.38  0.65 0.04
##            ppv_oob  0.00  0.19    0.24   0.27  0.27    0.29  0.34  0.45 0.05
##              npv_b  0.71  0.75    0.77   0.80  0.80    0.82  0.87  1.00 0.04
##            npv_oob  0.50  0.68    0.73   0.76  0.76    0.79  0.83  1.00 0.05
##  NAs
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0

# ROC plot
plot(PER.LYM.IMI)

# Youden index plot
plot_metric(PER.LYM.IMI)

# MAC IMI cow
MAC.IMI <- cutpointr(Pc.by.cow.C5, log.TOTAL.MAC.CO.COW, IMI.cow, 
                     method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(MAC.IMI)

## Method: maximize_metric 
## Predictor: log.TOTAL.MAC.CO.COW 
## Outcome: IMI.cow 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC   n n_pos n_neg
##  0.5733 419   311   108
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv
##            2.3026        1.1323 0.5847      0.6045      0.5278 0.7866 0.3167
##   tp  fn fp tn
##  188 123 51 57
## 
## Predictor summary: 
##     Data Min.       5%  1st Qu.   Median     Mean  3rd Qu.      95%     Max.
##  Overall    0 1.098612 1.945910 2.397895 2.672747 3.178054 5.651974 7.611842
##        0    0 1.098612 1.791759 2.197225 2.471720 2.833213 5.630198 7.611842
##        1    0 1.098612 1.945910 2.484907 2.742557 3.258097 5.579908 7.431892
##        SD NAs
##  1.335131   0
##  1.300397   0
##  1.341996   0
## 
## Bootstrap summary: 
##           Variable  Min.    5% 1st Qu. Median Mean 3rd Qu.  95% Max.   SD NAs
##   optimal_cutpoint  1.10  2.20    2.30   2.56 2.63    2.89 3.22 4.13 0.41   0
##              AUC_b  0.47  0.52    0.55   0.57 0.57    0.59 0.63 0.68 0.03   0
##            AUC_oob  0.42  0.51    0.55   0.57 0.57    0.60 0.64 0.69 0.04   0
##    sum_sens_spec_b  1.02  1.09    1.13   1.16 1.16    1.19 1.24 1.32 0.05   0
##  sum_sens_spec_oob  0.88  0.97    1.04   1.08 1.08    1.13 1.18 1.32 0.06   0
##              acc_b  0.32  0.41    0.47   0.54 0.54    0.60 0.64 0.75 0.08   0
##            acc_oob  0.30  0.38    0.45   0.51 0.51    0.57 0.62 0.74 0.08   0
##      sensitivity_b  0.12  0.26    0.35   0.49 0.49    0.62 0.69 0.97 0.15   0
##    sensitivity_oob  0.08  0.24    0.34   0.47 0.47    0.61 0.68 0.97 0.15   0
##      specificity_b  0.06  0.46    0.55   0.68 0.67    0.80 0.88 0.98 0.15   0
##    specificity_oob  0.02  0.36    0.49   0.60 0.61    0.75 0.86 0.97 0.17   0
##     cohens_kappa_b  0.02  0.06    0.09   0.12 0.12    0.15 0.19 0.26 0.04   0
##   cohens_kappa_oob -0.13 -0.02    0.03   0.06 0.06    0.09 0.14 0.23 0.05   0
##              ppv_b  0.71  0.76    0.79   0.82 0.82    0.84 0.88 0.96 0.04   0
##            ppv_oob  0.62  0.72    0.75   0.78 0.78    0.81 0.86 0.95 0.04   0
##              npv_b  0.22  0.26    0.29   0.32 0.32    0.34 0.38 0.59 0.04   0
##            npv_oob  0.06  0.22    0.26   0.29 0.29    0.31 0.35 0.47 0.04   0

# ROC plot
plot(MAC.IMI)

# Youden index plot
plot_metric(MAC.IMI)

# %MAC Major IMI cow
PER.MAC.IMI<- cutpointr(Pc.by.cow.C5, PERCENT.MAC.CO.COW, IMI.cow, 
                        method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(PER.MAC.IMI)

## Method: maximize_metric 
## Predictor: PERCENT.MAC.CO.COW 
## Outcome: IMI.cow 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC   n n_pos n_neg
##  0.5265 419   311   108
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv
##             15.56        1.0895 0.6158      0.6913      0.3981 0.7679 0.3094
##   tp fn fp tn
##  215 96 65 43
## 
## Predictor summary: 
##     Data Min.    5% 1st Qu. Median     Mean 3rd Qu.    95%  Max.       SD NAs
##  Overall 2.00 8.100  14.210 21.210 23.01516 29.8950 45.593 71.88 11.66827   0
##        0 2.13 6.292  13.405 20.125 22.43806 29.6325 45.375 59.38 12.13685   0
##        1 2.00 8.305  14.355 21.670 23.21556 30.8800 44.820 71.88 11.51440   0
## 
## Bootstrap summary: 
##           Variable  Min.    5% 1st Qu. Median  Mean 3rd Qu.   95%  Max.   SD
##   optimal_cutpoint  5.56  6.38   14.29  15.65 19.13   22.92 29.90 48.65 6.64
##              AUC_b  0.41  0.47    0.50   0.53  0.52    0.55  0.58  0.63 0.03
##            AUC_oob  0.36  0.46    0.50   0.53  0.53    0.56  0.60  0.68 0.04
##    sum_sens_spec_b  1.00  1.04    1.08   1.11  1.12    1.15  1.19  1.29 0.04
##  sum_sens_spec_oob  0.73  0.92    0.99   1.03  1.03    1.07  1.14  1.25 0.06
##              acc_b  0.24  0.41    0.52   0.60  0.58    0.65  0.74  0.82 0.10
##            acc_oob  0.21  0.37    0.48   0.56  0.55    0.62  0.71  0.82 0.10
##      sensitivity_b  0.04  0.27    0.48   0.66  0.61    0.75  0.98  1.00 0.20
##    sensitivity_oob  0.02  0.24    0.45   0.62  0.58    0.73  0.97  1.00 0.20
##      specificity_b  0.02  0.10    0.37   0.49  0.51    0.65  0.83  0.99 0.20
##    specificity_oob  0.00  0.06    0.30   0.42  0.44    0.59  0.79  0.98 0.21
##     cohens_kappa_b  0.00  0.04    0.07   0.10  0.10    0.12  0.17  0.28 0.04
##   cohens_kappa_oob -0.20 -0.07   -0.01   0.02  0.02    0.06  0.11  0.24 0.05
##              ppv_b  0.70  0.74    0.77   0.79  0.79    0.81  0.84  0.96 0.03
##            ppv_oob  0.50  0.69    0.73   0.75  0.75    0.78  0.81  0.91 0.04
##              npv_b  0.21  0.26    0.29   0.32  0.34    0.35  0.50  1.00 0.10
##            npv_oob  0.00  0.19    0.24   0.27  0.28    0.31  0.38  1.00 0.09
##  NAs
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    2

# ROC plot
plot(PER.MAC.IMI)

# Youden index plot
plot_metric(PER.MAC.IMI)

# Summary of results when using C5 and IMI as reference

print(c("Optimal SCC:", exp(SCC.IMI$optimal_cutpoint), SCC.IMI$sensitivity, SCC.IMI$specificity, SCC.IMI$ppv, SCC.IMI$npv, SCC.IMI$AUC))

##                                                 sensitivity         specificity 
##      "Optimal SCC:"               "122" "0.275080906148867"  "0.87962962962963" 
##                                                             
##  "0.86734693877551" "0.297805642633229" "0.546745774901115"

print(c("Optimal n NEU:", exp(NEU.IMI$optimal_cutpoint), NEU.IMI$sensitivity, NEU.IMI$specificity, NEU.IMI$ppv, NEU.IMI$npv, NEU.IMI$AUC))

##                                                 sensitivity         specificity 
##    "Optimal n NEU:"                "68" "0.247588424437299" "0.898148148148148" 
##                                                             
##             "0.875" "0.293051359516616"  "0.53685840181017"

print(c("Optimal %NEU:", PER.NEU.IMI$optimal_cutpoint, PER.NEU.IMI$sensitivity, PER.NEU.IMI$specificity, PER.NEU.IMI$ppv, PER.NEU.IMI$npv, PER.NEU.IMI$AUC))

##                                                 sensitivity         specificity 
##     "Optimal %NEU:"             "54.94" "0.559485530546624" "0.509259259259259" 
##                                                             
## "0.766519823788546" "0.286458333333333"  "0.50372156722639"

print(c("Optimal n LYM:", exp(LYM.IMI$optimal_cutpoint), LYM.IMI$sensitivity, LYM.IMI$specificity, LYM.IMI$ppv, LYM.IMI$npv, LYM.IMI$AUC))

##                                                 sensitivity         specificity 
##    "Optimal n LYM:"                "21" "0.279742765273312"  "0.87962962962963" 
##                                                             
##              "0.87" "0.297805642633229" "0.531499345004168"

print(c("Optimal %LYM:", PER.LYM.IMI$optimal_cutpoint, PER.LYM.IMI$sensitivity, PER.LYM.IMI$specificity, PER.LYM.IMI$ppv, PER.LYM.IMI$npv, PER.LYM.IMI$AUC))

##                                                 sensitivity         specificity 
##     "Optimal %LYM:"             "21.82" "0.481481481481481" "0.610932475884244" 
##                                                             
## "0.300578034682081" "0.772357723577236" "0.536173633440514"

print(c("Optimal n MAC:", exp(MAC.IMI$optimal_cutpoint), MAC.IMI$sensitivity, MAC.IMI$specificity, MAC.IMI$ppv, MAC.IMI$npv, MAC.IMI$AUC))

##                                                 sensitivity         specificity 
##    "Optimal n MAC:"                "10" "0.604501607717042" "0.527777777777778" 
##                                                             
## "0.786610878661088" "0.316666666666667" "0.573329760628796"

print(c("Optimal %MAC:", PER.MAC.IMI$optimal_cutpoint, PER.MAC.IMI$sensitivity, PER.MAC.IMI$specificity, PER.MAC.IMI$ppv, PER.MAC.IMI$npv, PER.MAC.IMI$AUC))

##                                                 sensitivity         specificity 
##     "Optimal %MAC:"             "15.56" "0.691318327974277" "0.398148148148148" 
##                                                             
## "0.767857142857143" "0.309352517985612" "0.526482672382994"

# CI using bootstrapping
# SCC.IMI
SCC.IMI_optimal_cutpoint <- boot_ci(SCC.IMI, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
SCC.IMI_lower_ci <- subset(SCC.IMI_optimal_cutpoint, SCC.IMI_optimal_cutpoint$quantile == 0.025)$values
exp(SCC.IMI_lower_ci)

## [1] 56

SCC.IMI_upper_ci <- subset(SCC.IMI_optimal_cutpoint, SCC.IMI_optimal_cutpoint$quantile == 0.975)$values
exp(SCC.IMI_upper_ci)

## [1] 154.0497

boot_ci(SCC.IMI, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.199
## 2    0.975  0.672

boot_ci(SCC.IMI, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.5  
## 2    0.975  0.946

boot_ci(SCC.IMI, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.781
## 2    0.975  0.935

boot_ci(SCC.IMI, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.251
## 2    0.975  0.365

boot_ci(SCC.IMI, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.486
## 2    0.975  0.604

# NEU.IMI
NEU.IMI_optimal_cutpoint <- boot_ci(NEU.IMI, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
NEU.IMI_lower_ci <- subset(NEU.IMI_optimal_cutpoint, NEU.IMI_optimal_cutpoint$quantile == 0.025)$values
exp(NEU.IMI_lower_ci)

## [1] 27

NEU.IMI_upper_ci <- subset(NEU.IMI_optimal_cutpoint, NEU.IMI_optimal_cutpoint$quantile == 0.975)$values
exp(NEU.IMI_upper_ci)

## [1] 76

boot_ci(NEU.IMI, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.212
## 2    0.975  0.608

boot_ci(NEU.IMI, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.545
## 2    0.975  0.946

boot_ci(NEU.IMI, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.777
## 2    0.975  0.938

boot_ci(NEU.IMI, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.251
## 2    0.975  0.361

boot_ci(NEU.IMI, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.479
## 2    0.975  0.594

# PER.NEU.IMI
PER.NEU.IMI_optimal_cutpoint <- boot_ci(PER.NEU.IMI, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
PER.NEU.IMI_optimal_cutpoint

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025   32.5
## 2    0.975   73.1

PER.NEU.IMI_lower_ci <- subset(PER.NEU.IMI_optimal_cutpoint, PER.NEU.IMI_optimal_cutpoint$quantile == 0.025)$values
PER.NEU.IMI_upper_ci <- subset(PER.NEU.IMI_optimal_cutpoint, PER.NEU.IMI_optimal_cutpoint$quantile == 0.975)$values

boot_ci(PER.NEU.IMI, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025 0.0502
## 2    0.975 0.987

boot_ci(PER.NEU.IMI, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025 0.0535
## 2    0.975 0.982

boot_ci(PER.NEU.IMI, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.726
## 2    0.975  0.917

boot_ci(PER.NEU.IMI, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.239
## 2    0.975  0.5

boot_ci(PER.NEU.IMI, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.438
## 2    0.975  0.568

# LYM.IMI
LYM.IMI_optimal_cutpoint <- boot_ci(LYM.IMI, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
LYM.IMI_lower_ci <- subset(LYM.IMI_optimal_cutpoint, LYM.IMI_optimal_cutpoint$quantile == 0.025)$values
exp(LYM.IMI_lower_ci)

## [1] 16

LYM.IMI_upper_ci <- subset(LYM.IMI_optimal_cutpoint, LYM.IMI_optimal_cutpoint$quantile == 0.975)$values
exp(LYM.IMI_upper_ci)

## [1] 23

boot_ci(LYM.IMI, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.231
## 2    0.975  0.400

boot_ci(LYM.IMI, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.768
## 2    0.975  0.941

boot_ci(LYM.IMI, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.803
## 2    0.975  0.934

boot_ci(LYM.IMI, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.25 
## 2    0.975  0.351

boot_ci(LYM.IMI, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.475
## 2    0.975  0.591

# PER.LYM.IMI
PER.LYM.IMI_optimal_cutpoint <- boot_ci(PER.LYM.IMI, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
PER.LYM.IMI_optimal_cutpoint

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025   15.3
## 2    0.975   29.1

PER.LYM.IMI_lower_ci <- subset(PER.LYM.IMI_optimal_cutpoint, PER.LYM.IMI_optimal_cutpoint$quantile == 0.025)$values
PER.LYM.IMI_upper_ci <- subset(PER.LYM.IMI_optimal_cutpoint, PER.LYM.IMI_optimal_cutpoint$quantile == 0.975)$values

boot_ci(PER.LYM.IMI, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.186
## 2    0.975  0.918

boot_ci(PER.LYM.IMI, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.208
## 2    0.975  0.894

boot_ci(PER.LYM.IMI, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.245
## 2    0.975  0.406

boot_ci(PER.LYM.IMI, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.738
## 2    0.975  0.889

boot_ci(PER.LYM.IMI, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.471
## 2    0.975  0.599

# MAC.IMI
MAC.IMI_optimal_cutpoint <- boot_ci(MAC.IMI, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
MAC.IMI_lower_ci <- subset(MAC.IMI_optimal_cutpoint, MAC.IMI_optimal_cutpoint$quantile == 0.025)$values
exp(MAC.IMI_lower_ci)

## [1] 9

MAC.IMI_upper_ci <- subset(MAC.IMI_optimal_cutpoint, MAC.IMI_optimal_cutpoint$quantile == 0.975)$values
exp(MAC.IMI_upper_ci)

## [1] 29

boot_ci(MAC.IMI, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.237
## 2    0.975  0.711

boot_ci(MAC.IMI, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.425
## 2    0.975  0.903

boot_ci(MAC.IMI, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.752
## 2    0.975  0.894

boot_ci(MAC.IMI, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.256
## 2    0.975  0.392

boot_ci(MAC.IMI, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.510
## 2    0.975  0.634

# PER.MAC.IMI
PER.MAC.IMI_optimal_cutpoint <- boot_ci(PER.MAC.IMI, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
PER.MAC.IMI_optimal_cutpoint

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025   6.36
## 2    0.975  30

PER.MAC.IMI_lower_ci <- subset(PER.MAC.IMI_optimal_cutpoint, PER.MAC.IMI_optimal_cutpoint$quantile == 0.025)$values
PER.MAC.IMI_upper_ci <- subset(PER.MAC.IMI_optimal_cutpoint, PER.MAC.IMI_optimal_cutpoint$quantile == 0.975)$values

boot_ci(PER.MAC.IMI, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.254
## 2    0.975  0.987

boot_ci(PER.MAC.IMI, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025 0.0654
## 2    0.975 0.856

boot_ci(PER.MAC.IMI, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.731
## 2    0.975  0.853

boot_ci(PER.MAC.IMI, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.254
## 2    0.975  0.667

boot_ci(PER.MAC.IMI, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.458
## 2    0.975  0.587

# Prevalence

# True prevalence

table(Pc.by.cow.C5$IMI.cow)

## 
##   0   1 
## 108 313

prop.table(table(Pc.by.cow.C5$IMI.cow))

## 
##         0         1 
## 0.2565321 0.7434679

#Apparent prevalence

# ---- SCC ----

# optimal cut-off

optimal_scc_C5_IMI_cow <- SCC.IMI$optimal_cutpoint

table(Pc.by.cow.C5$SCC.CO.COW > exp(optimal_scc_C5_IMI_cow))

## 
## FALSE  TRUE 
##   322    98

prop.table(table(Pc.by.cow.C5$SCC.CO.COW > exp(optimal_scc_C5_IMI_cow)))

## 
##     FALSE      TRUE 
## 0.7666667 0.2333333

# confidence intervals

table(Pc.by.cow.C5$SCC.CO.COW > exp(SCC.IMI_lower_ci))

## 
## FALSE  TRUE 
##   168   252

prop.table(table(Pc.by.cow.C5$SCC.CO.COW > exp(SCC.IMI_lower_ci)))

## 
## FALSE  TRUE 
##   0.4   0.6

table(Pc.by.cow.C5$SCC.CO.COW > exp(SCC.IMI_upper_ci))

## 
## FALSE  TRUE 
##   342    78

prop.table(table(Pc.by.cow.C5$SCC.CO.COW >exp(SCC.IMI_upper_ci)))

## 
##     FALSE      TRUE 
## 0.8142857 0.1857143

# ---- Neutrophils ----

# optimal cut-off

optimal_neu_C5_IMI_cow <- NEU.IMI$optimal_cutpoint

table(Pc.by.cow.C5$TOTAL.NEU.CO.COW > exp(optimal_neu_C5_IMI_cow))

## 
## FALSE  TRUE 
##   334    88

prop.table(table(Pc.by.cow.C5$TOTAL.NEU.CO.COW > exp(optimal_neu_C5_IMI_cow)))

## 
##     FALSE      TRUE 
## 0.7914692 0.2085308

# confidence intervals

table(Pc.by.cow.C5$TOTAL.NEU.CO.COW > exp(NEU.IMI_lower_ci))

## 
## FALSE  TRUE 
##   196   226

prop.table(table(Pc.by.cow.C5$TOTAL.NEU.CO.COW > exp(NEU.IMI_lower_ci)))

## 
##    FALSE     TRUE 
## 0.464455 0.535545

table(Pc.by.cow.C5$TOTAL.NEU.CO.COW > exp(NEU.IMI_upper_ci))

## 
## FALSE  TRUE 
##   343    79

prop.table(table(Pc.by.cow.C5$TOTAL.NEU.CO.COW >exp(NEU.IMI_upper_ci)))

## 
##     FALSE      TRUE 
## 0.8127962 0.1872038

# ---- Percent neu ----

# Optimal cut-off

optimal_per_neu_C5_IMI_cow <-PER.NEU.IMI$optimal_cutpoint

table(Pc.by.cow.C5$PERCENT.NEU.CO.COW > optimal_per_neu_C5_IMI_cow)

## 
## FALSE  TRUE 
##   194   228

prop.table(table(Pc.by.cow.C5$PERCENT.NEU.CO.COW > optimal_per_neu_C5_IMI_cow))

## 
##     FALSE      TRUE 
## 0.4597156 0.5402844

# confidence intervals

table(Pc.by.cow.C5$PERCENT.NEU.CO.COW > PER.NEU.IMI_lower_ci)

## 
## FALSE  TRUE 
##     9   413

prop.table(table(Pc.by.cow.C5$PERCENT.NEU.CO.COW > PER.NEU.IMI_lower_ci))

## 
##      FALSE       TRUE 
## 0.02132701 0.97867299

table(Pc.by.cow.C5$PERCENT.NEU.CO.COW > PER.NEU.IMI_upper_ci)

## 
## FALSE  TRUE 
##   401    21

prop.table(table(Pc.by.cow.C5$PERCENT.NEU.CO.COW >PER.NEU.IMI_upper_ci))

## 
##      FALSE       TRUE 
## 0.95023697 0.04976303

# ---- Lymphocytes ----

# optimal cut-off

optimal_lym_C5_IMI_cow <-LYM.IMI$optimal_cutpoint

table(Pc.by.cow.C5$TOTAL.LYM.CO.COW > exp(optimal_lym_C5_IMI_cow))

## 
## FALSE  TRUE 
##   324    98

prop.table(table(Pc.by.cow.C5$TOTAL.LYM.CO.COW > exp(optimal_lym_C5_IMI_cow)))

## 
##     FALSE      TRUE 
## 0.7677725 0.2322275

# confidence intervals

table(Pc.by.cow.C5$TOTAL.LYM.CO.COW > exp(LYM.IMI_lower_ci))

## 
## FALSE  TRUE 
##   267   155

prop.table(table(Pc.by.cow.C5$TOTAL.LYM.CO.COW > exp(LYM.IMI_lower_ci)))

## 
##     FALSE      TRUE 
## 0.6327014 0.3672986

table(Pc.by.cow.C5$TOTAL.LYM.CO.COW > exp(LYM.IMI_upper_ci))

## 
## FALSE  TRUE 
##   330    92

prop.table(table(Pc.by.cow.C5$TOTAL.LYM.CO.COW >exp(LYM.IMI_upper_ci)))

## 
##     FALSE      TRUE 
## 0.7819905 0.2180095

# ---- Percent lym ----

# optimal cut-off

optimal_per_lym_C5_IMI_cow <-PER.LYM.IMI$optimal_cutpoint

table(Pc.by.cow.C5$PERCENT.LYM.CO.COW > optimal_per_lym_C5_IMI_cow)

## 
## FALSE  TRUE 
##   250   172

prop.table(table(Pc.by.cow.C5$PERCENT.LYM.CO.COW > optimal_per_lym_C5_IMI_cow))

## 
##     FALSE      TRUE 
## 0.5924171 0.4075829

# confidence intervals

table(Pc.by.cow.C5$PERCENT.LYM.CO.COW > PER.LYM.IMI_lower_ci)

## 
## FALSE  TRUE 
##    97   325

prop.table(table(Pc.by.cow.C5$PERCENT.LYM.CO.COW > PER.LYM.IMI_lower_ci))

## 
##     FALSE      TRUE 
## 0.2298578 0.7701422

table(Pc.by.cow.C5$PERCENT.LYM.CO.COW > PER.LYM.IMI_upper_ci)

## 
## FALSE  TRUE 
##   368    54

prop.table(table(Pc.by.cow.C5$PERCENT.LYM.CO.COW >PER.LYM.IMI_upper_ci))

## 
##     FALSE      TRUE 
## 0.8720379 0.1279621

# ---- Macrophages ----

# optimal cut-off

optimal_mac_C5_IMI_cow <- MAC.IMI$optimal_cutpoint

table(Pc.by.cow.C5$TOTAL.MAC.CO.COW > exp(optimal_mac_C5_IMI_cow))

## 
## FALSE  TRUE 
##   193   229

prop.table(table(Pc.by.cow.C5$TOTAL.MAC.CO.COW > exp(optimal_mac_C5_IMI_cow)))

## 
##    FALSE     TRUE 
## 0.457346 0.542654

# confidence intervals

table(Pc.by.cow.C5$TOTAL.MAC.CO.COW > exp(MAC.IMI_lower_ci))

## 
## FALSE  TRUE 
##   181   241

prop.table(table(Pc.by.cow.C5$TOTAL.MAC.CO.COW > exp(MAC.IMI_lower_ci)))

## 
##   FALSE    TRUE 
## 0.42891 0.57109

table(Pc.by.cow.C5$TOTAL.MAC.CO.COW > exp(MAC.IMI_upper_ci))

## 
## FALSE  TRUE 
##   333    89

prop.table(table(Pc.by.cow.C5$TOTAL.MAC.CO.COW >exp(MAC.IMI_upper_ci)))

## 
##     FALSE      TRUE 
## 0.7890995 0.2109005

# Percent MAC

# optimal cut-off

optimal_per_mac_C5_IMI_cow <- PER.MAC.IMI$optimal_cutpoint

table(Pc.by.cow.C5$PERCENT.MAC.CO.COW > optimal_per_mac_C5_IMI_cow)

## 
## FALSE  TRUE 
##   140   282

prop.table(table(Pc.by.cow.C5$PERCENT.MAC.CO.COW > optimal_per_mac_C5_IMI_cow))

## 
##     FALSE      TRUE 
## 0.3317536 0.6682464

# confidence intervals

table(Pc.by.cow.C5$PERCENT.MAC.CO.COW > PER.MAC.IMI_lower_ci)

## 
## FALSE  TRUE 
##    13   409

prop.table(table(Pc.by.cow.C5$PERCENT.MAC.CO.COW > PER.MAC.IMI_lower_ci))

## 
##      FALSE       TRUE 
## 0.03080569 0.96919431

table(Pc.by.cow.C5$PERCENT.MAC.CO.COW > PER.MAC.IMI_upper_ci)

## 
## FALSE  TRUE 
##   318   104

prop.table(table(Pc.by.cow.C5$PERCENT.MAC.CO.COW >PER.MAC.IMI_upper_ci))

## 
##     FALSE      TRUE 
## 0.7535545 0.2464455

Cow high SCC

# Set a seed for reproducibility

set.seed(123)

# Delaval SCM (SCC>200.000) reference

# n NEU SCM COW
NEU.SCM <- cutpointr(Pc.by.cow.C5, log.TOTAL.NEU.CO.COW, delaval.SCM.CO.COW, 
                            method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(NEU.SCM)

## Method: maximize_metric 
## Predictor: log.TOTAL.NEU.CO.COW 
## Outcome: delaval.SCM.CO.COW 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC   n n_pos n_neg
##  0.9726 420    63   357
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv tp
##            4.4427        1.8973 0.9571      0.9365      0.9608 0.8082 0.9885 59
##  fn fp  tn
##   4 14 343
## 
## Predictor summary: 
##     Data     Min.       5%  1st Qu.   Median     Mean  3rd Qu.      95%
##  Overall 1.386294 2.297317 2.890372 3.401197 3.674843 4.064717 6.232366
##        0 1.386294 2.079442 2.833213 3.218876 3.276409 3.761200 4.343805
##        1 2.833213 4.330004 4.979177 5.762051 5.932633 6.931957 8.484060
##      Max.        SD NAs
##  8.948976 1.2673930   0
##  8.350902 0.7143865   0
##  8.948976 1.3531688   0
## 
## Bootstrap summary: 
##           Variable Min.   5% 1st Qu. Median Mean 3rd Qu.  95% Max.   SD NAs
##   optimal_cutpoint 4.06 4.32    4.38   4.44 4.43    4.47 4.58 4.82 0.09   0
##              AUC_b 0.91 0.95    0.96   0.97 0.97    0.98 0.99 1.00 0.01   0
##            AUC_oob 0.91 0.94    0.96   0.97 0.97    0.99 1.00 1.00 0.02   0
##    sum_sens_spec_b 1.77 1.85    1.88   1.90 1.90    1.92 1.95 1.98 0.03   0
##  sum_sens_spec_oob 1.50 1.77    1.83   1.87 1.87    1.91 1.96 1.99 0.06   0
##              acc_b 0.87 0.93    0.95   0.96 0.96    0.97 0.98 1.00 0.01   0
##            acc_oob 0.84 0.92    0.94   0.95 0.95    0.96 0.97 0.99 0.02   0
##      sensitivity_b 0.83 0.89    0.93   0.95 0.94    0.97 0.99 1.00 0.03   0
##    sensitivity_oob 0.52 0.80    0.88   0.92 0.91    0.96 1.00 1.00 0.06   0
##      specificity_b 0.86 0.93    0.95   0.96 0.96    0.97 0.98 1.00 0.02   0
##    specificity_oob 0.81 0.91    0.94   0.96 0.95    0.97 0.98 1.00 0.02   0
##     cohens_kappa_b 0.57 0.76    0.81   0.84 0.84    0.87 0.91 0.98 0.05   0
##   cohens_kappa_oob 0.54 0.71    0.77   0.81 0.81    0.85 0.90 0.97 0.06   0
##              ppv_b 0.48 0.69    0.76   0.81 0.80    0.85 0.90 0.99 0.07   0
##            ppv_oob 0.44 0.65    0.73   0.79 0.79    0.85 0.91 1.00 0.08   0
##              npv_b 0.97 0.98    0.99   0.99 0.99    0.99 1.00 1.00 0.01   0
##            npv_oob 0.92 0.96    0.98   0.98 0.98    0.99 1.00 1.00 0.01   0

# ROC plot
plot(NEU.SCM)

# Youden index plot
plot_metric(NEU.SCM)

# %NEU SCM COW
PER.NEU.SCM <- cutpointr(Pc.by.cow.C5, PERCENT.NEU.CO.COW, delaval.SCM.CO.COW, 
                            method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(PER.NEU.SCM)

## Method: maximize_metric 
## Predictor: PERCENT.NEU.CO.COW 
## Outcome: delaval.SCM.CO.COW 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC   n n_pos n_neg
##  0.5992 420    63   357
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity   ppv    npv tp
##             62.21         1.226 0.6976      0.4921      0.7339 0.246 0.8912 31
##  fn fp  tn
##  32 95 262
## 
## Predictor summary: 
##     Data  Min.     5% 1st Qu. Median     Mean 3rd Qu.    95%  Max.       SD NAs
##  Overall 18.75 37.090  49.425 56.475 56.18600  64.355 73.088 86.73 11.07392   0
##        0 18.75 37.500  49.450 55.560 55.67866  62.790 72.444 86.73 10.69669   0
##        1 27.33 34.938  49.960 61.370 59.06095  69.330 75.299 81.93 12.72340   0
## 
## Bootstrap summary: 
##           Variable  Min.    5% 1st Qu. Median  Mean 3rd Qu.   95%  Max.   SD
##   optimal_cutpoint 43.23 56.32   61.14  62.21 63.44   66.75 68.95 72.27 3.77
##              AUC_b  0.43  0.53    0.57   0.60  0.60    0.63  0.67  0.75 0.04
##            AUC_oob  0.36  0.50    0.56   0.60  0.60    0.64  0.69  0.80 0.06
##    sum_sens_spec_b  1.03  1.15    1.20   1.25  1.25    1.29  1.35  1.52 0.06
##  sum_sens_spec_oob  0.80  1.01    1.10   1.16  1.15    1.21  1.30  1.46 0.09
##              acc_b  0.23  0.57    0.68   0.71  0.72    0.79  0.83  0.87 0.08
##            acc_oob  0.22  0.54    0.65   0.69  0.70    0.76  0.82  0.87 0.08
##      sensitivity_b  0.13  0.28    0.39   0.50  0.49    0.57  0.69  0.98 0.13
##    sensitivity_oob  0.00  0.19    0.30   0.41  0.41    0.50  0.62  0.87 0.14
##      specificity_b  0.12  0.55    0.70   0.74  0.76    0.86  0.92  0.96 0.11
##    specificity_oob  0.11  0.53    0.68   0.74  0.75    0.85  0.91  0.97 0.12
##     cohens_kappa_b  0.03  0.10    0.15   0.19  0.19    0.23  0.29  0.39 0.06
##   cohens_kappa_oob -0.13  0.01    0.07   0.12  0.12    0.17  0.25  0.36 0.07
##              ppv_b  0.14  0.19    0.23   0.27  0.28    0.32  0.43  0.53 0.07
##            ppv_oob  0.00  0.14    0.19   0.23  0.24    0.27  0.37  0.56 0.07
##              npv_b  0.82  0.86    0.88   0.89  0.89    0.91  0.93  0.98 0.02
##            npv_oob  0.76  0.83    0.86   0.88  0.88    0.89  0.92  0.97 0.03
##  NAs
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0

# ROC plot
plot(PER.NEU.SCM)

# Youden index plot
plot_metric(PER.NEU.SCM)

# Lym
LYM.SCM  <- cutpointr(Pc.by.cow.C5, log.TOTAL.LYM.CO.COW, delaval.SCM.CO.COW, 
                         method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(LYM.SCM)

## Method: maximize_metric 
## Predictor: log.TOTAL.LYM.CO.COW 
## Outcome: delaval.SCM.CO.COW 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC   n n_pos n_neg
##  0.9769 420    63   357
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv tp
##            3.1355         1.873 0.9143      0.9683      0.9048 0.6421 0.9938 61
##  fn fp  tn
##   2 34 323
## 
## Predictor summary: 
##     Data     Min.       5%  1st Qu.   Median     Mean  3rd Qu.      95%
##  Overall 0.000000 1.098612 1.945910 2.484907 2.633699 2.944439 5.136961
##        0 0.000000 1.098612 1.791759 2.302585 2.273239 2.772589 3.401197
##        1 2.564949 3.258097 3.939267 4.442651 4.676303 5.280659 6.418765
##      Max.        SD NAs
##  7.123673 1.1794695   0
##  6.562444 0.7513396   0
##  7.123673 1.0827760   0
## 
## Bootstrap summary: 
##           Variable Min.   5% 1st Qu. Median Mean 3rd Qu.  95% Max.   SD NAs
##   optimal_cutpoint 3.14 3.14    3.14   3.26 3.26    3.26 3.56 3.87 0.16   0
##              AUC_b 0.94 0.96    0.97   0.98 0.98    0.98 0.99 1.00 0.01   0
##            AUC_oob 0.93 0.96    0.97   0.98 0.98    0.99 0.99 1.00 0.01   0
##    sum_sens_spec_b 1.77 1.83    1.86   1.88 1.88    1.90 1.92 1.97 0.03   0
##  sum_sens_spec_oob 1.57 1.74    1.81   1.85 1.84    1.88 1.92 1.96 0.06   0
##              acc_b 0.87 0.90    0.91   0.92 0.93    0.94 0.97 0.98 0.02   0
##            acc_oob 0.85 0.89    0.90   0.92 0.92    0.93 0.95 0.98 0.02   0
##      sensitivity_b 0.82 0.90    0.94   0.96 0.96    0.98 1.00 1.00 0.03   0
##    sensitivity_oob 0.60 0.78    0.89   0.95 0.92    0.96 1.00 1.00 0.07   0
##      specificity_b 0.86 0.89    0.90   0.92 0.92    0.94 0.98 0.99 0.03   0
##    specificity_oob 0.84 0.87    0.90   0.92 0.92    0.94 0.98 1.00 0.03   0
##     cohens_kappa_b 0.58 0.67    0.71   0.75 0.76    0.80 0.87 0.93 0.06   0
##   cohens_kappa_oob 0.52 0.63    0.69   0.73 0.73    0.77 0.83 0.93 0.06   0
##              ppv_b 0.49 0.58    0.63   0.68 0.69    0.74 0.87 0.95 0.09   0
##            ppv_oob 0.43 0.54    0.61   0.67 0.68    0.73 0.86 1.00 0.10   0
##              npv_b 0.97 0.98    0.99   0.99 0.99    1.00 1.00 1.00 0.01   0
##            npv_oob 0.93 0.96    0.98   0.99 0.99    0.99 1.00 1.00 0.01   0

# ROC plot
plot(LYM.SCM)

# Youden index plot
plot_metric(LYM.SCM)

# %LYM SCM COW
PER.LYM.SCM <- cutpointr(Pc.by.cow.C5, PERCENT.LYM.CO.COW, delaval.SCM.CO.COW, 
                            method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 0

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(PER.LYM.SCM)

## Method: maximize_metric 
## Predictor: PERCENT.LYM.CO.COW 
## Outcome: delaval.SCM.CO.COW 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC   n n_pos n_neg
##  0.6827 420   357    63
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv
##             21.79         1.394 0.5405      0.4734      0.9206 0.9713 0.2358
##   tp  fn fp tn
##  169 188  5 58
## 
## Predictor summary: 
##     Data Min.      5% 1st Qu. Median     Mean 3rd Qu.    95%  Max.       SD NAs
##  Overall 0.00 10.4485 15.7175  19.96 20.77726 25.0575 34.292 44.79 7.242575   0
##        0 0.00 10.5220 16.4600  21.05 21.43543 26.3200 34.620 44.79 7.438397   0
##        1 6.55  9.1440 14.8000  17.02 17.04762 19.9550 25.020 27.78 4.490512   0
## 
## Bootstrap summary: 
##           Variable  Min.    5% 1st Qu. Median  Mean 3rd Qu.   95%  Max.   SD
##   optimal_cutpoint 16.67 21.05   21.79  21.79 21.63   21.82 21.88 26.32 0.75
##              AUC_b  0.57  0.63    0.66   0.68  0.68    0.70  0.73  0.78 0.03
##            AUC_oob  0.54  0.61    0.65   0.68  0.68    0.71  0.75  0.83 0.04
##    sum_sens_spec_b  1.21  1.33    1.37   1.40  1.40    1.43  1.46  1.53 0.04
##  sum_sens_spec_oob  0.95  1.20    1.32   1.37  1.36    1.42  1.48  1.61 0.08
##              acc_b  0.33  0.50    0.53   0.55  0.55    0.56  0.59  0.75 0.04
##            acc_oob  0.36  0.48    0.51   0.54  0.54    0.56  0.60  0.71 0.04
##      sensitivity_b  0.24  0.43    0.46   0.48  0.48    0.50  0.54  0.78 0.05
##    sensitivity_oob  0.25  0.41    0.45   0.47  0.48    0.50  0.55  0.78 0.05
##      specificity_b  0.57  0.85    0.90   0.92  0.92    0.95  0.97  1.00 0.05
##    specificity_oob  0.26  0.70    0.86   0.90  0.88    0.95  1.00  1.00 0.11
##     cohens_kappa_b  0.07  0.14    0.17   0.18  0.18    0.20  0.23  0.33 0.03
##   cohens_kappa_oob -0.03  0.09    0.14   0.16  0.16    0.19  0.23  0.33 0.04
##              ppv_b  0.89  0.94    0.96   0.97  0.97    0.98  0.99  1.00 0.01
##            ppv_oob  0.81  0.90    0.95   0.97  0.96    0.98  1.00  1.00 0.03
##              npv_b  0.14  0.19    0.22   0.24  0.24    0.26  0.29  0.40 0.03
##            npv_oob  0.10  0.17    0.20   0.23  0.23    0.25  0.29  0.35 0.04
##  NAs
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0

# ROC plot
plot(PER.LYM.SCM)

# Youden index plot
plot_metric(PER.LYM.SCM)

# MAC SCM COW
MAC.SCM <- cutpointr(Pc.by.cow.C5, log.TOTAL.MAC.CO.COW, delaval.SCM.CO.COW, 
                        method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(MAC.SCM)

## Method: maximize_metric 
## Predictor: log.TOTAL.MAC.CO.COW 
## Outcome: delaval.SCM.CO.COW 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC   n n_pos n_neg
##  0.9595 420    63   357
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv tp
##            3.2958        1.8067 0.9024      0.9048       0.902 0.6196 0.9817 57
##  fn fp  tn
##   6 35 322
## 
## Predictor summary: 
##     Data     Min.       5%  1st Qu.   Median     Mean  3rd Qu.      95%
##  Overall 0.000000 1.098612 1.945910 2.397895 2.668517 3.178054 5.527306
##        0 0.000000 1.098612 1.791759 2.302585 2.271053 2.772589 3.668625
##        1 2.079442 2.949568 3.960405 4.859812 4.920813 6.041444 7.142157
##      Max.        SD NAs
##  7.611842 1.3313114   0
##  7.611842 0.8558791   0
##  7.431892 1.3067079   0
## 
## Bootstrap summary: 
##           Variable Min.   5% 1st Qu. Median Mean 3rd Qu.  95% Max.   SD NAs
##   optimal_cutpoint 2.89 2.89    3.30   3.52 3.51    3.76 3.89 4.14 0.29   0
##              AUC_b 0.90 0.93    0.95   0.96 0.96    0.97 0.98 1.00 0.01   0
##            AUC_oob 0.89 0.93    0.95   0.96 0.96    0.98 0.99 1.00 0.02   0
##    sum_sens_spec_b 1.69 1.76    1.80   1.82 1.82    1.85 1.89 1.94 0.04   0
##  sum_sens_spec_oob 1.30 1.65    1.72   1.77 1.76    1.81 1.86 1.94 0.07   0
##              acc_b 0.79 0.85    0.90   0.93 0.92    0.95 0.96 0.98 0.03   0
##            acc_oob 0.76 0.83    0.89   0.91 0.91    0.93 0.95 0.98 0.04   0
##      sensitivity_b 0.73 0.82    0.86   0.90 0.90    0.93 0.97 1.00 0.05   0
##    sensitivity_oob 0.32 0.71    0.79   0.85 0.85    0.91 0.96 1.00 0.08   0
##      specificity_b 0.75 0.83    0.90   0.93 0.92    0.96 0.98 1.00 0.04   0
##    specificity_oob 0.71 0.81    0.89   0.93 0.92    0.96 0.98 0.99 0.05   0
##     cohens_kappa_b 0.42 0.57    0.68   0.74 0.73    0.80 0.85 0.91 0.09   0
##   cohens_kappa_oob 0.35 0.51    0.63   0.69 0.68    0.75 0.82 0.90 0.09   0
##              ppv_b 0.36 0.49    0.62   0.70 0.70    0.80 0.87 0.98 0.12   0
##            ppv_oob 0.30 0.46    0.58   0.67 0.67    0.77 0.87 0.96 0.13   0
##              npv_b 0.95 0.97    0.98   0.98 0.98    0.99 0.99 1.00 0.01   0
##            npv_oob 0.90 0.94    0.96   0.97 0.97    0.98 0.99 1.00 0.01   0

# ROC plot
plot(MAC.SCM)

# Youden index plot
plot_metric(MAC.SCM)

# %MAC SCM COW
PER.MAC.SCM <- cutpointr(Pc.by.cow.C5, PERCENT.MAC.CO.COW, delaval.SCM.CO.COW, 
                            method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 0

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(PER.MAC.SCM)

## Method: maximize_metric 
## Predictor: PERCENT.MAC.CO.COW 
## Outcome: delaval.SCM.CO.COW 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC   n n_pos n_neg
##  0.4753 420   357    63
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv
##             21.67        1.0299 0.4976      0.4902      0.5397 0.8578 0.1574
##   tp  fn fp tn
##  175 182 29 34
## 
## Predictor summary: 
##     Data Min.    5% 1st Qu. Median     Mean 3rd Qu.     95%  Max.       SD NAs
##  Overall 2.00 8.120   14.25  21.19 23.03700 29.9175 45.5565 71.88 11.66521   0
##        0 2.00 8.000   14.13  21.21 22.88622 29.7900 44.5460 71.88 11.62919   0
##        1 5.94 8.862   14.83  21.16 23.89143 32.1950 45.3430 52.07 11.92548   0
## 
## Bootstrap summary: 
##           Variable  Min.    5% 1st Qu. Median  Mean 3rd Qu.   95%  Max.    SD
##   optimal_cutpoint  6.00  9.16   12.88  21.38 23.24   24.76 52.15 54.24 12.74
##              AUC_b  0.34  0.41    0.45   0.48  0.48    0.50  0.54  0.60  0.04
##            AUC_oob  0.26  0.39    0.44   0.47  0.47    0.51  0.57  0.64  0.05
##    sum_sens_spec_b  0.82  1.02    1.05   1.08  1.08    1.11  1.16  1.27  0.05
##  sum_sens_spec_oob  0.62  0.83    0.91   0.96  0.95    1.00  1.04  1.17  0.07
##              acc_b  0.13  0.17    0.45   0.53  0.53    0.74  0.81  0.89  0.20
##            acc_oob  0.12  0.16    0.41   0.49  0.50    0.70  0.79  0.90  0.20
##      sensitivity_b  0.01  0.03    0.42   0.51  0.53    0.82  0.93  0.99  0.28
##    sensitivity_oob  0.00  0.02    0.39   0.49  0.51    0.80  0.92  0.99  0.29
##      specificity_b  0.03  0.12    0.27   0.59  0.55    0.69  1.00  1.00  0.28
##    specificity_oob  0.00  0.04    0.15   0.45  0.44    0.58  1.00  1.00  0.30
##     cohens_kappa_b -0.10  0.01    0.03   0.05  0.05    0.07  0.12  0.22  0.03
##   cohens_kappa_oob -0.18 -0.11   -0.06  -0.02 -0.03    0.00  0.03  0.13  0.04
##              ppv_b  0.71  0.84    0.86   0.88  0.89    0.90  1.00  1.00  0.05
##            ppv_oob  0.40  0.75    0.81   0.84  0.84    0.86  1.00  1.00  0.07
##              npv_b  0.08  0.14    0.16   0.18  0.19    0.20  0.27  0.50  0.05
##            npv_oob  0.00  0.05    0.11   0.13  0.13    0.16  0.19  0.27  0.04
##  NAs
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##   14
##    0
##    0

# ROC plot
plot(PER.MAC.SCM)

# Youden index plot
plot_metric(PER.MAC.SCM)

# Summary of results when using SCM as reference

print(c("Optimal n NEU:", exp(NEU.SCM$optimal_cutpoint), NEU.SCM$sensitivity, NEU.SCM$specificity, NEU.SCM$ppv, NEU.SCM$npv, NEU.SCM$AUC))

##                                                 sensitivity         specificity 
##    "Optimal n NEU:"                "85" "0.936507936507937"  "0.96078431372549" 
##                                                             
## "0.808219178082192" "0.988472622478386" "0.972633497843582"

print(c("Optimal %NEU:", PER.NEU.SCM$optimal_cutpoint, PER.NEU.SCM$sensitivity, PER.NEU.SCM$specificity, PER.NEU.SCM$ppv, PER.NEU.SCM$npv, PER.NEU.SCM$AUC))

##                                                 sensitivity         specificity 
##     "Optimal %NEU:"             "62.21" "0.492063492063492" "0.733893557422969" 
##                                                             
## "0.246031746031746" "0.891156462585034" "0.599217464763683"

print(c("Optimal n LYM:", exp(LYM.SCM$optimal_cutpoint), LYM.SCM$sensitivity, LYM.SCM$specificity, LYM.SCM$ppv, LYM.SCM$npv, LYM.SCM$AUC))

##                                                 sensitivity         specificity 
##    "Optimal n LYM:"                "23" "0.968253968253968" "0.904761904761905" 
##                                                             
## "0.642105263157895" "0.993846153846154" "0.976901871859855"

print(c("Optimal %LYM:", PER.LYM.SCM$optimal_cutpoint, PER.LYM.SCM$sensitivity, PER.LYM.SCM$specificity, PER.LYM.SCM$ppv, PER.LYM.SCM$npv, PER.LYM.SCM$AUC))

##                                                 sensitivity         specificity 
##     "Optimal %LYM:"             "21.79" "0.473389355742297" "0.920634920634921" 
##                                                             
## "0.971264367816092" "0.235772357723577" "0.682739762571695"

print(c("Optimal n MAC:", exp(MAC.SCM$optimal_cutpoint), MAC.SCM$sensitivity, MAC.SCM$specificity, MAC.SCM$ppv, MAC.SCM$npv, MAC.SCM$AUC))

##                                                 sensitivity         specificity 
##    "Optimal n MAC:"                "27" "0.904761904761905" "0.901960784313726" 
##                                                             
## "0.619565217391304" "0.981707317073171" "0.959539371304077"

print(c("Optimal %MAC:", PER.MAC.SCM$optimal_cutpoint, PER.MAC.SCM$sensitivity, PER.MAC.SCM$specificity, PER.MAC.SCM$ppv, PER.MAC.SCM$npv, PER.MAC.SCM$AUC))

##                                                 sensitivity         specificity 
##     "Optimal %MAC:"             "21.67" "0.490196078431373"  "0.53968253968254" 
##                                                             
## "0.857843137254902" "0.157407407407407" "0.475301231603753"

# CI using bootstrapping
# NEU.SCM
NEU.SCM_optimal_cutpoint <- boot_ci(NEU.SCM, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
NEU.SCM_lower_ci <- subset(NEU.SCM_optimal_cutpoint, NEU.SCM_optimal_cutpoint$quantile == 0.025)$values
exp(NEU.SCM_lower_ci)

## [1] 75

NEU.SCM_upper_ci <- subset(NEU.SCM_optimal_cutpoint, NEU.SCM_optimal_cutpoint$quantile == 0.975)$values
exp(NEU.SCM_upper_ci)

## [1] 98

boot_ci(NEU.SCM, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.881
## 2    0.975  1

boot_ci(NEU.SCM, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.923
## 2    0.975  0.986

boot_ci(NEU.SCM, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.667
## 2    0.975  0.922

boot_ci(NEU.SCM, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.977
## 2    0.975  1

boot_ci(NEU.SCM, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.941
## 2    0.975  0.995

# PER.NEU.SCM
PER.NEU.SCM_optimal_cutpoint <- boot_ci(PER.NEU.SCM, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
PER.NEU.SCM_optimal_cutpoint

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025   56.3
## 2    0.975   69.2

PER.NEU.SCM_lower_ci <- subset(PER.NEU.SCM_optimal_cutpoint, PER.NEU.SCM_optimal_cutpoint$quantile == 0.025)$values
PER.NEU.SCM_upper_ci <- subset(PER.NEU.SCM_optimal_cutpoint, PER.NEU.SCM_optimal_cutpoint$quantile == 0.975)$values

boot_ci(PER.NEU.SCM, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.254
## 2    0.975  0.721

boot_ci(PER.NEU.SCM, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.528
## 2    0.975  0.928

boot_ci(PER.NEU.SCM, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.184
## 2    0.975  0.452

boot_ci(PER.NEU.SCM, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.858
## 2    0.975  0.933

boot_ci(PER.NEU.SCM, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.512
## 2    0.975  0.686

# LYM.SCM
LYM.SCM_optimal_cutpoint <- boot_ci(LYM.SCM, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
LYM.SCM_lower_ci <- subset(LYM.SCM_optimal_cutpoint, LYM.SCM_optimal_cutpoint$quantile == 0.025)$values
exp(LYM.SCM_lower_ci)

## [1] 23

LYM.SCM_upper_ci <- subset(LYM.SCM_optimal_cutpoint, LYM.SCM_optimal_cutpoint$quantile == 0.975)$values
exp(LYM.SCM_upper_ci)

## [1] 35

boot_ci(LYM.SCM, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.884
## 2    0.975  1

boot_ci(LYM.SCM, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.883
## 2    0.975  0.980

boot_ci(LYM.SCM, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.563
## 2    0.975  0.885

boot_ci(LYM.SCM, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.979
## 2    0.975  1

boot_ci(LYM.SCM, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.959
## 2    0.975  0.990

# PER.LYM.SCM
PER.LYM.SCM_optimal_cutpoint <- boot_ci(PER.LYM.SCM, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
PER.LYM.SCM_optimal_cutpoint

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025   19.0
## 2    0.975   21.9

PER.LYM.SCM_lower_ci <- subset(PER.LYM.SCM_optimal_cutpoint, PER.LYM.SCM_optimal_cutpoint$quantile == 0.025)$values
PER.LYM.SCM_upper_ci <- subset(PER.LYM.SCM_optimal_cutpoint, PER.LYM.SCM_optimal_cutpoint$quantile == 0.975)$values

boot_ci(PER.LYM.SCM, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.423
## 2    0.975  0.646

boot_ci(PER.LYM.SCM, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.768
## 2    0.975  0.983

boot_ci(PER.LYM.SCM, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.937
## 2    0.975  0.994

boot_ci(PER.LYM.SCM, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.185
## 2    0.975  0.302

boot_ci(PER.LYM.SCM, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.621
## 2    0.975  0.740

# MAC.SCM
MAC.SCM_optimal_cutpoint <- boot_ci(MAC.SCM, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
MAC.SCM_lower_ci <- subset(MAC.SCM_optimal_cutpoint, MAC.SCM_optimal_cutpoint$quantile == 0.025)$values
exp(MAC.SCM_lower_ci)

## [1] 18

MAC.SCM_upper_ci <- subset(MAC.SCM_optimal_cutpoint, MAC.SCM_optimal_cutpoint$quantile == 0.975)$values
exp(MAC.SCM_upper_ci)

## [1] 50

boot_ci(MAC.SCM, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.806
## 2    0.975  0.984

boot_ci(MAC.SCM, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.802
## 2    0.975  0.981

boot_ci(MAC.SCM, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.455
## 2    0.975  0.892

boot_ci(MAC.SCM, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.963
## 2    0.975  0.997

boot_ci(MAC.SCM, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.929
## 2    0.975  0.985

# PER.MAC.SCM
PER.MAC.SCM_optimal_cutpoint <- boot_ci(PER.MAC.SCM, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
PER.MAC.SCM_optimal_cutpoint

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025   9.09
## 2    0.975  53.2

PER.MAC.SCM_lower_ci <- subset(PER.MAC.SCM_optimal_cutpoint, PER.MAC.SCM_optimal_cutpoint$quantile == 0.025)$values
PER.MAC.SCM_upper_ci <- subset(PER.MAC.SCM_optimal_cutpoint, PER.MAC.SCM_optimal_cutpoint$quantile == 0.975)$values

boot_ci(PER.MAC.SCM, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025 0.0197
## 2    0.975 0.948

boot_ci(PER.MAC.SCM, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025 0.0986
## 2    0.975 1

boot_ci(PER.MAC.SCM, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.833
## 2    0.975  1

boot_ci(PER.MAC.SCM, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.129
## 2    0.975  0.3

boot_ci(PER.MAC.SCM, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.397
## 2    0.975  0.555

# Prevalence

# True prevalence

table(Pc.by.cow.C5$delaval.SCM.CO.COW)

## 
##   0   1 
## 357  63

prop.table(table(Pc.by.cow.C5$delaval.SCM.CO.COW))

## 
##    0    1 
## 0.85 0.15

#Apparent prevalence

# ---- Neutrophils ----

# optimal cut-off

optimal_neu_C5_SCM_cow <- NEU.SCM$optimal_cutpoint

table(Pc.by.cow.C5$TOTAL.NEU.CO.COW > exp(optimal_neu_C5_SCM_cow))

## 
## FALSE  TRUE 
##   349    73

prop.table(table(Pc.by.cow.C5$TOTAL.NEU.CO.COW > exp(optimal_neu_C5_SCM_cow)))

## 
##     FALSE      TRUE 
## 0.8270142 0.1729858

# confidence intervals

table(Pc.by.cow.C5$TOTAL.NEU.CO.COW > exp(NEU.SCM_lower_ci))

## 
## FALSE  TRUE 
##   338    84

prop.table(table(Pc.by.cow.C5$TOTAL.NEU.CO.COW > exp(NEU.SCM_lower_ci)))

## 
##     FALSE      TRUE 
## 0.8009479 0.1990521

table(Pc.by.cow.C5$TOTAL.NEU.CO.COW > exp(NEU.SCM_upper_ci))

## 
## FALSE  TRUE 
##   359    63

prop.table(table(Pc.by.cow.C5$TOTAL.NEU.CO.COW >exp(NEU.SCM_upper_ci)))

## 
##     FALSE      TRUE 
## 0.8507109 0.1492891

# ---- Percent neu ----

# Optimal cut-off

optimal_per_neu_C5_SCM_cow <-PER.NEU.SCM$optimal_cutpoint

table(Pc.by.cow.C5$PERCENT.NEU.CO.COW > optimal_per_neu_C5_SCM_cow)

## 
## FALSE  TRUE 
##   297   125

prop.table(table(Pc.by.cow.C5$PERCENT.NEU.CO.COW > optimal_per_neu_C5_SCM_cow))

## 
##     FALSE      TRUE 
## 0.7037915 0.2962085

# confidence intervals

table(Pc.by.cow.C5$PERCENT.NEU.CO.COW > PER.NEU.SCM_lower_ci)

## 
## FALSE  TRUE 
##   210   212

prop.table(table(Pc.by.cow.C5$PERCENT.NEU.CO.COW > PER.NEU.SCM_lower_ci))

## 
##     FALSE      TRUE 
## 0.4976303 0.5023697

table(Pc.by.cow.C5$PERCENT.NEU.CO.COW > PER.NEU.SCM_upper_ci)

## 
## FALSE  TRUE 
##   373    49

prop.table(table(Pc.by.cow.C5$PERCENT.NEU.CO.COW >PER.NEU.SCM_upper_ci))

## 
##     FALSE      TRUE 
## 0.8838863 0.1161137

# ---- Lymphocytes ----

# optimal cut-off

optimal_lym_C5_SCM_cow <-LYM.SCM$optimal_cutpoint

table(Pc.by.cow.C5$TOTAL.LYM.CO.COW > exp(optimal_lym_C5_SCM_cow))

## 
## FALSE  TRUE 
##   330    92

prop.table(table(Pc.by.cow.C5$TOTAL.LYM.CO.COW > exp(optimal_lym_C5_SCM_cow)))

## 
##     FALSE      TRUE 
## 0.7819905 0.2180095

# confidence intervals

table(Pc.by.cow.C5$TOTAL.LYM.CO.COW > exp(LYM.SCM_lower_ci))

## 
## FALSE  TRUE 
##   330    92

prop.table(table(Pc.by.cow.C5$TOTAL.LYM.CO.COW > exp(LYM.SCM_lower_ci)))

## 
##     FALSE      TRUE 
## 0.7819905 0.2180095

table(Pc.by.cow.C5$TOTAL.LYM.CO.COW > exp(LYM.SCM_upper_ci))

## 
## FALSE  TRUE 
##   355    67

prop.table(table(Pc.by.cow.C5$TOTAL.LYM.CO.COW >exp(LYM.SCM_upper_ci)))

## 
##     FALSE      TRUE 
## 0.8412322 0.1587678

# ---- Percent lym ----

# optimal cut-off

optimal_per_lym_C5_SCM_cow <-PER.LYM.SCM$optimal_cutpoint

table(Pc.by.cow.C5$PERCENT.LYM.CO.COW > optimal_per_lym_C5_SCM_cow)

## 
## FALSE  TRUE 
##   248   174

prop.table(table(Pc.by.cow.C5$PERCENT.LYM.CO.COW > optimal_per_lym_C5_SCM_cow))

## 
##     FALSE      TRUE 
## 0.5876777 0.4123223

# confidence intervals

table(Pc.by.cow.C5$PERCENT.LYM.CO.COW > PER.LYM.SCM_lower_ci)

## 
## FALSE  TRUE 
##   187   235

prop.table(table(Pc.by.cow.C5$PERCENT.LYM.CO.COW > PER.LYM.SCM_lower_ci))

## 
##    FALSE     TRUE 
## 0.443128 0.556872

table(Pc.by.cow.C5$PERCENT.LYM.CO.COW > PER.LYM.SCM_upper_ci)

## 
## FALSE  TRUE 
##   252   170

prop.table(table(Pc.by.cow.C5$PERCENT.LYM.CO.COW >PER.LYM.SCM_upper_ci))

## 
##     FALSE      TRUE 
## 0.5971564 0.4028436

# ---- Macrophages ----

# optimal cut-off

optimal_mac_C5_SCM_cow <- MAC.SCM$optimal_cutpoint

table(Pc.by.cow.C5$TOTAL.MAC.CO.COW > exp(optimal_mac_C5_SCM_cow))

## 
## FALSE  TRUE 
##   331    91

prop.table(table(Pc.by.cow.C5$TOTAL.MAC.CO.COW > exp(optimal_mac_C5_SCM_cow)))

## 
##     FALSE      TRUE 
## 0.7843602 0.2156398

# confidence intervals

table(Pc.by.cow.C5$TOTAL.MAC.CO.COW > exp(MAC.SCM_lower_ci))

## 
## FALSE  TRUE 
##   288   134

prop.table(table(Pc.by.cow.C5$TOTAL.MAC.CO.COW > exp(MAC.SCM_lower_ci)))

## 
##     FALSE      TRUE 
## 0.6824645 0.3175355

table(Pc.by.cow.C5$TOTAL.MAC.CO.COW > exp(MAC.SCM_upper_ci))

## 
## FALSE  TRUE 
##   362    60

prop.table(table(Pc.by.cow.C5$TOTAL.MAC.CO.COW >exp(MAC.SCM_upper_ci)))

## 
##     FALSE      TRUE 
## 0.8578199 0.1421801

# Percent MAC

# optimal cut-off

optimal_per_mac_C5_SCM_cow <- PER.MAC.SCM$optimal_cutpoint

table(Pc.by.cow.C5$PERCENT.MAC.CO.COW > optimal_per_mac_C5_SCM_cow)

## 
## FALSE  TRUE 
##   218   204

prop.table(table(Pc.by.cow.C5$PERCENT.MAC.CO.COW > optimal_per_mac_C5_SCM_cow))

## 
##     FALSE      TRUE 
## 0.5165877 0.4834123

# confidence intervals

table(Pc.by.cow.C5$PERCENT.MAC.CO.COW > PER.MAC.SCM_lower_ci)

## 
## FALSE  TRUE 
##    31   391

prop.table(table(Pc.by.cow.C5$PERCENT.MAC.CO.COW > PER.MAC.SCM_lower_ci))

## 
##      FALSE       TRUE 
## 0.07345972 0.92654028

table(Pc.by.cow.C5$PERCENT.MAC.CO.COW > PER.MAC.SCM_upper_ci)

## 
## FALSE  TRUE 
##   416     6

prop.table(table(Pc.by.cow.C5$PERCENT.MAC.CO.COW >PER.MAC.SCM_upper_ci))

## 
##      FALSE       TRUE 
## 0.98578199 0.01421801

Quarter IMI

# Set a seed for reproducibility

set.seed(123)

# IMI QUARTER referent

# SCC IMI quarter

SCC.IMI <- cutpointr(Pc.by.quarter.C5, log.SCC, IMI, 
                     method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(SCC.IMI)

## Method: maximize_metric 
## Predictor: log.SCC 
## Outcome: IMI 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC    n n_pos n_neg
##  0.5151 1552   538  1014
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv
##            4.9628        1.0725 0.6392      0.2007      0.8718 0.4538 0.6728
##   tp  fn  fp  tn
##  108 430 130 884
## 
## Predictor summary: 
##     Data     Min.       5%  1st Qu.   Median     Mean  3rd Qu.      95%
##  Overall 1.386294 2.772589 3.526361 4.043051 4.160545 4.590018 6.203706
##        0 1.386294 2.811995 3.526361 4.043051 4.113878 4.553877 5.666724
##        1 1.791759 2.708050 3.526361 4.051747 4.248501 4.781162 6.899994
##      Max.        SD NAs
##  8.522181 1.0520387   0
##  8.488176 0.9515683   0
##  8.522181 1.2152049   0
## 
## Bootstrap summary: 
##           Variable  Min.   5% 1st Qu. Median Mean 3rd Qu.  95% Max.   SD NAs
##   optimal_cutpoint  3.53 4.47    4.76   4.84 4.87    4.96 5.29 7.52 0.27   0
##              AUC_b  0.47 0.49    0.50   0.52 0.52    0.53 0.54 0.56 0.02   0
##            AUC_oob  0.43 0.48    0.50   0.52 0.52    0.53 0.55 0.58 0.02   0
##    sum_sens_spec_b  1.02 1.05    1.07   1.08 1.08    1.09 1.11 1.14 0.02   0
##  sum_sens_spec_oob  0.92 1.01    1.04   1.06 1.05    1.07 1.10 1.15 0.03   0
##              acc_b  0.45 0.60    0.62   0.63 0.63    0.64 0.66 0.70 0.02   0
##            acc_oob  0.40 0.57    0.61   0.62 0.62    0.64 0.66 0.69 0.03   0
##      sensitivity_b  0.05 0.14    0.21   0.23 0.24    0.27 0.36 0.78 0.06   0
##    sensitivity_oob  0.04 0.12    0.19   0.22 0.22    0.25 0.33 0.74 0.06   0
##      specificity_b  0.27 0.72    0.82   0.85 0.84    0.87 0.93 0.99 0.06   0
##    specificity_oob  0.21 0.70    0.81   0.84 0.83    0.87 0.92 0.99 0.06   0
##     cohens_kappa_b  0.03 0.06    0.08   0.09 0.09    0.11 0.13 0.16 0.02   0
##   cohens_kappa_oob -0.07 0.01    0.04   0.06 0.06    0.08 0.11 0.17 0.03   0
##              ppv_b  0.36 0.40    0.42   0.45 0.45    0.47 0.52 0.74 0.04   0
##            ppv_oob  0.28 0.35    0.39   0.42 0.42    0.45 0.50 0.74 0.05   0
##              npv_b  0.63 0.65    0.67   0.68 0.68    0.69 0.70 0.72 0.01   0
##            npv_oob  0.58 0.64    0.66   0.67 0.67    0.68 0.70 0.72 0.02   0

# ROC plot

plot(SCC.IMI)

# Youden index plot

plot_metric(SCC.IMI)

# NEU IMI quarter

NEU.IMI <- cutpointr(Pc.by.quarter.C5, log.Neu, IMI, 
                     method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(NEU.IMI)

## Method: maximize_metric 
## Predictor: log.Neu 
## Outcome: IMI 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC    n n_pos n_neg
##  0.5251 1556   541  1015
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv
##            3.9318          1.08 0.6234      0.2662      0.8138 0.4324 0.6754
##   tp  fn  fp  tn
##  144 397 189 826
## 
## Predictor summary: 
##     Data     Min.       5%  1st Qu.   Median     Mean  3rd Qu.      95%
##  Overall 1.098612 2.197225 2.833213 3.258097 3.435124 3.806662 5.330222
##        0 1.098612 2.197225 2.833213 3.218876 3.369005 3.761200 4.996741
##        1 1.098612 2.197225 2.833213 3.295837 3.559175 3.970292 6.226537
##       Max.        SD NAs
##  10.174773 1.1008755   0
##   9.036463 0.9590292   0
##  10.174773 1.3187892   0
## 
## Bootstrap summary: 
##           Variable  Min.   5% 1st Qu. Median Mean 3rd Qu.  95% Max.   SD NAs
##   optimal_cutpoint  2.56 3.69    3.93   3.93 4.04    4.11 4.71 6.04 0.31   0
##              AUC_b  0.47 0.50    0.51   0.52 0.52    0.54 0.55 0.59 0.02   0
##            AUC_oob  0.46 0.49    0.51   0.53 0.53    0.54 0.56 0.59 0.02   0
##    sum_sens_spec_b  1.02 1.05    1.07   1.08 1.08    1.10 1.12 1.16 0.02   0
##  sum_sens_spec_oob  0.92 1.01    1.04   1.06 1.06    1.08 1.11 1.15 0.03   0
##              acc_b  0.41 0.60    0.62   0.63 0.63    0.64 0.66 0.69 0.02   0
##            acc_oob  0.36 0.57    0.61   0.62 0.62    0.64 0.66 0.68 0.03   0
##      sensitivity_b  0.05 0.13    0.23   0.26 0.25    0.28 0.36 0.90 0.07   0
##    sensitivity_oob  0.04 0.11    0.21   0.24 0.24    0.27 0.32 0.84 0.07   0
##      specificity_b  0.15 0.73    0.81   0.83 0.83    0.86 0.94 0.98 0.07   0
##    specificity_oob  0.12 0.71    0.80   0.82 0.82    0.85 0.94 0.99 0.07   0
##     cohens_kappa_b  0.03 0.06    0.08   0.09 0.09    0.11 0.13 0.17 0.02   0
##   cohens_kappa_oob -0.07 0.01    0.05   0.07 0.07    0.09 0.12 0.16 0.03   0
##              ppv_b  0.35 0.40    0.43   0.45 0.45    0.47 0.52 0.69 0.04   0
##            ppv_oob  0.29 0.35    0.40   0.42 0.43    0.45 0.50 0.75 0.05   0
##              npv_b  0.63 0.65    0.67   0.68 0.68    0.69 0.70 0.74 0.01   0
##            npv_oob  0.58 0.64    0.66   0.67 0.67    0.68 0.70 0.74 0.02   0

# ROC plot
plot(NEU.IMI)

# Youden index plot
plot_metric(NEU.IMI)

# %NEU IMI quarter
PER.NEU.IMI <- cutpointr(Pc.by.quarter.C5, PERCENT01_N, IMI, 
                         method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(PER.NEU.IMI)

## Method: maximize_metric 
## Predictor: PERCENT01_N 
## Outcome: IMI 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC    n n_pos n_neg
##  0.5189 1556   541  1015
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv
##             60.51        1.0438 0.5694       0.366      0.6778 0.3771 0.6673
##   tp  fn  fp  tn
##  198 343 327 688
## 
## Predictor summary: 
##     Data  Min.      5% 1st Qu. Median     Mean 3rd Qu.     95%  Max.       SD
##  Overall 13.79 38.3825  50.000 56.285 56.30482   63.27 73.7025 89.55 10.65339
##        0 21.89 38.1740  49.415 56.100 56.04310   62.96 72.8050 89.55 10.54335
##        1 13.79 38.6100  50.000 56.760 56.79584   64.00 74.3600 88.07 10.84965
##  NAs
##    0
##    0
##    0
## 
## Bootstrap summary: 
##           Variable  Min.    5% 1st Qu. Median  Mean 3rd Qu.   95%  Max.   SD
##   optimal_cutpoint 28.05 48.48   51.79  60.34 58.11   60.83 68.03 76.06 6.96
##              AUC_b  0.47  0.49    0.51   0.52  0.52    0.53  0.54  0.57 0.02
##            AUC_oob  0.44  0.48    0.50   0.52  0.52    0.53  0.55  0.59 0.02
##    sum_sens_spec_b  1.01  1.03    1.04   1.06  1.06    1.07  1.09  1.14 0.02
##  sum_sens_spec_oob  0.91  0.96    0.99   1.00  1.01    1.02  1.05  1.12 0.03
##              acc_b  0.34  0.43    0.48   0.57  0.54    0.60  0.64  0.68 0.07
##            acc_oob  0.34  0.40    0.45   0.55  0.52    0.58  0.62  0.67 0.07
##      sensitivity_b  0.04  0.16    0.35   0.40  0.48    0.71  0.83  1.00 0.23
##    sensitivity_oob  0.02  0.12    0.30   0.37  0.45    0.67  0.80  0.99 0.23
##      specificity_b  0.01  0.23    0.35   0.67  0.58    0.71  0.89  0.98 0.23
##    specificity_oob  0.00  0.21    0.33   0.66  0.56    0.70  0.87  0.97 0.23
##     cohens_kappa_b  0.00  0.02    0.04   0.06  0.06    0.07  0.09  0.14 0.02
##   cohens_kappa_oob -0.08 -0.04   -0.01   0.00  0.01    0.02  0.05  0.12 0.03
##              ppv_b  0.32  0.35    0.37   0.38  0.39    0.40  0.45  0.62 0.03
##            ppv_oob  0.19  0.31    0.33   0.35  0.35    0.37  0.40  0.54 0.03
##              npv_b  0.63  0.65    0.67   0.68  0.68    0.69  0.72  1.00 0.02
##            npv_oob  0.33  0.61    0.64   0.66  0.65    0.67  0.69  0.74 0.03
##  NAs
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0

# ROC plot
plot(PER.NEU.IMI)

# Youden index plot
plot_metric(PER.NEU.IMI)

# LYM IMI quarter

LYM.IMI <-  cutpointr(Pc.by.quarter.C5, log.Lym, IMI, 
                      method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(LYM.IMI)

## Method: maximize_metric 
## Predictor: log.Lym 
## Outcome: IMI 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC    n n_pos n_neg
##  0.5289 1556   541  1015
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv
##            3.2189        1.0822 0.6395      0.2181       0.864 0.4609 0.6746
##   tp  fn  fp  tn
##  118 423 138 877
## 
## Predictor summary: 
##     Data Min.       5%  1st Qu.   Median     Mean  3rd Qu.      95%     Max.
##  Overall    0 1.098612 1.791759 2.302585 2.436436 2.890372 4.330733 8.165932
##        0    0 1.098612 1.791759 2.302585 2.380298 2.833213 4.012739 7.264730
##        1    0 1.098612 1.791759 2.397895 2.541760 3.044522 5.003946 8.165932
##         SD NAs
##  1.0779589   0
##  0.9744082   0
##  1.2434561   0
## 
## Bootstrap summary: 
##           Variable  Min.   5% 1st Qu. Median Mean 3rd Qu.  95% Max.   SD NAs
##   optimal_cutpoint  2.20 2.48    2.77   3.22 3.06    3.22 3.30 4.87 0.29   0
##              AUC_b  0.48 0.50    0.52   0.53 0.53    0.54 0.55 0.58 0.02   0
##            AUC_oob  0.46 0.50    0.51   0.53 0.53    0.54 0.56 0.60 0.02   0
##    sum_sens_spec_b  1.02 1.05    1.07   1.09 1.09    1.10 1.12 1.16 0.02   0
##  sum_sens_spec_oob  0.94 1.00    1.04   1.06 1.06    1.09 1.11 1.17 0.03   0
##              acc_b  0.49 0.57    0.61   0.63 0.62    0.65 0.66 0.68 0.03   0
##            acc_oob  0.44 0.53    0.59   0.63 0.61    0.64 0.66 0.70 0.04   0
##      sensitivity_b  0.07 0.19    0.22   0.24 0.28    0.35 0.47 0.67 0.10   0
##    sensitivity_oob  0.04 0.17    0.21   0.23 0.26    0.31 0.44 0.65 0.09   0
##      specificity_b  0.41 0.62    0.74   0.86 0.81    0.87 0.88 0.98 0.10   0
##    specificity_oob  0.34 0.59    0.72   0.85 0.80    0.87 0.89 0.98 0.10   0
##     cohens_kappa_b  0.02 0.06    0.08   0.09 0.10    0.11 0.13 0.17 0.02   0
##   cohens_kappa_oob -0.05 0.00    0.04   0.07 0.07    0.10 0.13 0.19 0.04   0
##              ppv_b  0.35 0.38    0.42   0.45 0.45    0.48 0.51 0.68 0.04   0
##            ppv_oob  0.29 0.34    0.39   0.43 0.43    0.46 0.51 0.57 0.05   0
##              npv_b  0.63 0.66    0.67   0.68 0.68    0.69 0.70 0.73 0.01   0
##            npv_oob  0.59 0.64    0.66   0.67 0.67    0.68 0.70 0.73 0.02   0

# ROC plot
plot(LYM.IMI)

# Youden index plot
plot_metric(LYM.IMI)

# %LYM IMI quarter
PER.LYM.IMI <- cutpointr(Pc.by.quarter.C5, PERCENT02_N, IMI, 
                         method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(PER.LYM.IMI)

## Method: maximize_metric 
## Predictor: PERCENT02_N 
## Outcome: IMI 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##  AUC    n n_pos n_neg
##  0.5 1556   541  1015
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv
##             22.12         1.035 0.5276      0.4843      0.5507 0.3649 0.6671
##   tp  fn  fp  tn
##  262 279 456 559
## 
## Predictor summary: 
##     Data Min.      5% 1st Qu. Median     Mean 3rd Qu.    95%  Max.       SD NAs
##  Overall 0.00  9.8375 16.1525 21.325 21.92774 27.0925 36.640 61.54 8.136529   0
##        0 4.17 10.0700 16.2650 21.170 21.98306 27.1800 36.642 61.54 8.071498   0
##        1 0.00  9.1500 16.0000 21.670 21.82396 27.0300 36.170 50.82 8.263679   0
## 
## Bootstrap summary: 
##           Variable  Min.    5% 1st Qu. Median Mean 3rd Qu.   95% Max.   SD NAs
##   optimal_cutpoint  7.90 19.72   22.12  22.92  Inf   25.00 36.92  Inf  NaN   0
##              AUC_b  0.44  0.47    0.49   0.50 0.50    0.51  0.53 0.56 0.02   0
##            AUC_oob  0.43  0.47    0.49   0.50 0.50    0.51  0.53 0.56 0.02   0
##    sum_sens_spec_b  1.00  1.01    1.03   1.05 1.05    1.06  1.09 1.14 0.02   0
##  sum_sens_spec_oob  0.89  0.95    0.99   1.01 1.01    1.03  1.06 1.11 0.03   0
##              acc_b  0.35  0.49    0.53   0.55 0.55    0.57  0.63 0.68 0.04   0
##            acc_oob  0.33  0.47    0.51   0.53 0.53    0.55  0.62 0.68 0.04   0
##      sensitivity_b  0.00  0.06    0.37   0.46 0.44    0.51  0.61 0.98 0.14   0
##    sensitivity_oob  0.00  0.04    0.34   0.43 0.41    0.48  0.58 0.96 0.14   0
##      specificity_b  0.02  0.44    0.55   0.60 0.61    0.69  0.96 1.00 0.14   0
##    specificity_oob  0.02  0.41    0.53   0.58 0.59    0.67  0.95 1.00 0.14   0
##     cohens_kappa_b  0.00  0.01    0.03   0.04 0.05    0.06  0.08 0.13 0.02   0
##   cohens_kappa_oob -0.12 -0.04   -0.01   0.01 0.01    0.03  0.05 0.10 0.03   0
##              ppv_b  0.32  0.35    0.36   0.38 0.38    0.39  0.42 1.00 0.04   2
##            ppv_oob  0.00  0.31    0.33   0.35 0.35    0.37  0.39 0.54 0.04   7
##              npv_b  0.62  0.64    0.66   0.67 0.67    0.68  0.70 0.76 0.02   0
##            npv_oob  0.41  0.62    0.64   0.65 0.65    0.67  0.69 0.74 0.02   0

# ROC plot
plot(PER.LYM.IMI)

## Warning: Removed 2 rows containing non-finite outside the scale range
## (`stat_density()`).

# Youden index plot
plot_metric(PER.LYM.IMI)

# MAC IMI quarter
MAC.IMI <- cutpointr(Pc.by.quarter.C5, log.Mac, IMI, 
                     method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(MAC.IMI)

## Method: maximize_metric 
## Predictor: log.Mac 
## Outcome: IMI 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC    n n_pos n_neg
##  0.5176 1556   541  1015
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv tp
##            3.4657         1.057 0.6433      0.1516      0.9054 0.4607 0.6669 82
##   fn fp  tn
##  459 96 919
## 
## Predictor summary: 
##     Data Min.        5%  1st Qu.   Median     Mean  3rd Qu.      95%     Max.
##  Overall    0 0.6931472 1.609438 2.197225 2.351310 2.787745 4.521789 9.127176
##        0    0 1.0986123 1.609438 2.197225 2.310454 2.772589 4.334655 8.196988
##        1    0 0.6931472 1.609438 2.302585 2.427963 2.944439 4.962845 9.127176
##        SD NAs
##  1.208201   0
##  1.093028   0
##  1.396821   0
## 
## Bootstrap summary: 
##           Variable  Min.    5% 1st Qu. Median Mean 3rd Qu.  95% Max.   SD NAs
##   optimal_cutpoint  1.95  2.40    2.71   3.04 3.05    3.47 3.53 7.46 0.44   0
##              AUC_b  0.47  0.49    0.51   0.52 0.52    0.53 0.54 0.58 0.02   0
##            AUC_oob  0.43  0.48    0.50   0.52 0.52    0.53 0.55 0.58 0.02   0
##    sum_sens_spec_b  1.01  1.04    1.06   1.07 1.07    1.08 1.10 1.13 0.02   0
##  sum_sens_spec_oob  0.89  0.98    1.01   1.04 1.03    1.05 1.08 1.13 0.03   0
##              acc_b  0.47  0.54    0.60   0.63 0.62    0.64 0.66 0.68 0.04   0
##            acc_oob  0.42  0.51    0.58   0.61 0.60    0.63 0.66 0.68 0.04   0
##      sensitivity_b  0.02  0.13    0.17   0.24 0.27    0.34 0.51 0.72 0.12   0
##    sensitivity_oob  0.00  0.12    0.15   0.21 0.24    0.30 0.47 0.65 0.11   0
##      specificity_b  0.33  0.57    0.73   0.83 0.80    0.90 0.92 1.00 0.12   0
##    specificity_oob  0.31  0.54    0.72   0.82 0.79    0.90 0.92 1.00 0.12   0
##     cohens_kappa_b  0.02  0.04    0.06   0.08 0.08    0.09 0.11 0.14 0.02   0
##   cohens_kappa_oob -0.10 -0.02    0.02   0.04 0.04    0.06 0.09 0.15 0.03   0
##              ppv_b  0.33  0.37    0.40   0.43 0.44    0.46 0.51 0.86 0.05   0
##            ppv_oob  0.00  0.33    0.36   0.39 0.40    0.43 0.49 0.89 0.05   0
##              npv_b  0.63  0.65    0.66   0.67 0.67    0.68 0.70 0.72 0.01   0
##            npv_oob  0.60  0.63    0.65   0.66 0.66    0.67 0.69 0.72 0.02   0

# ROC plot
plot(MAC.IMI)

# Youden index plot
plot_metric(MAC.IMI)

# %MAC Major IMI cow
PER.MAC.IMI<- cutpointr(Pc.by.quarter.C5, PERCENT03_N, IMI.cow, 
                        method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 0

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(PER.MAC.IMI)

## Method: maximize_metric 
## Predictor: PERCENT03_N 
## Outcome: IMI.cow 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC    n n_pos n_neg
##  0.5237 1635   417  1218
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv
##             26.92        1.0492 0.6171      0.3357      0.7135 0.2863 0.7583
##   tp  fn  fp  tn
##  140 277 349 869
## 
## Predictor summary: 
##     Data Min.     5% 1st Qu. Median     Mean 3rd Qu.    95%  Max.       SD NAs
##  Overall 0.00 6.4200 13.0400   20.0 21.77527  28.615 43.270 69.57 11.52476   0
##        0 2.16 7.5400 13.7900   20.5 22.43103  29.670 43.568 60.39 11.35891   0
##        1 0.00 6.2365 12.7675   20.0 21.55076  28.330 42.860 69.57 11.57708   0
## 
## Bootstrap summary: 
##           Variable  Min.    5% 1st Qu. Median  Mean 3rd Qu.   95%  Max.   SD
##   optimal_cutpoint  5.83  8.74   11.76  20.17 20.04   27.08 30.92 41.94 8.60
##              AUC_b  0.47  0.50    0.51   0.52  0.52    0.53  0.55  0.58 0.02
##            AUC_oob  0.45  0.49    0.51   0.52  0.52    0.54  0.56  0.59 0.02
##    sum_sens_spec_b  1.01  1.04    1.05   1.07  1.07    1.08  1.10  1.14 0.02
##  sum_sens_spec_oob  0.84  0.96    0.99   1.01  1.01    1.03  1.06  1.11 0.03
##              acc_b  0.28  0.32    0.38   0.52  0.51    0.63  0.68  0.74 0.13
##            acc_oob  0.26  0.31    0.36   0.49  0.49    0.61  0.66  0.74 0.13
##      sensitivity_b  0.07  0.24    0.34   0.56  0.58    0.84  0.95  0.99 0.27
##    sensitivity_oob  0.04  0.18    0.29   0.48  0.54    0.81  0.93  0.97 0.27
##      specificity_b  0.04  0.11    0.22   0.51  0.49    0.73  0.82  0.96 0.27
##    specificity_oob  0.04  0.10    0.21   0.49  0.47    0.72  0.82  0.95 0.27
##     cohens_kappa_b  0.01  0.02    0.04   0.05  0.05    0.07  0.09  0.14 0.02
##   cohens_kappa_oob -0.13 -0.04   -0.01   0.01  0.01    0.02  0.05  0.10 0.02
##              ppv_b  0.24  0.25    0.27   0.29  0.29    0.31  0.33  0.44 0.03
##            ppv_oob  0.17  0.22    0.25   0.26  0.26    0.27  0.30  0.35 0.02
##              npv_b  0.73  0.75    0.76   0.78  0.79    0.81  0.86  0.92 0.03
##            npv_oob  0.66  0.72    0.74   0.75  0.76    0.77  0.82  0.91 0.03
##  NAs
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0

# ROC plot
plot(PER.MAC.IMI)

# Youden index plot
plot_metric(PER.MAC.IMI)

# Summary of results when using C5 and IMI as the reference

print(c("Optimal SCC:", exp(SCC.IMI$optimal_cutpoint), SCC.IMI$sensitivity, SCC.IMI$specificity, SCC.IMI$ppv, SCC.IMI$npv, SCC.IMI$AUC))

##                                                 sensitivity         specificity 
##      "Optimal SCC:"               "143" "0.200743494423792" "0.871794871794872" 
##                                                             
## "0.453781512605042" "0.672754946727549" "0.515092606849827"

print(c("Optimal n NEU:", exp(NEU.IMI$optimal_cutpoint), NEU.IMI$sensitivity, NEU.IMI$specificity, NEU.IMI$ppv, NEU.IMI$npv, NEU.IMI$AUC))

##                                                 sensitivity         specificity 
##    "Optimal n NEU:"                "51" "0.266173752310536" "0.813793103448276" 
##                                                             
## "0.432432432432432" "0.675388389206868" "0.525104941587828"

print(c("Optimal %NEU:", PER.NEU.IMI$optimal_cutpoint, PER.NEU.IMI$sensitivity, PER.NEU.IMI$specificity, PER.NEU.IMI$ppv, PER.NEU.IMI$npv, PER.NEU.IMI$AUC))

##                                                 sensitivity         specificity 
##     "Optimal %NEU:"             "60.51" "0.365988909426987" "0.677832512315271" 
##                                                             
## "0.377142857142857" "0.667313288069835" "0.518874006355681"

print(c("Optimal n LYM:", exp(LYM.IMI$optimal_cutpoint), LYM.IMI$sensitivity, LYM.IMI$specificity, LYM.IMI$ppv, LYM.IMI$npv, LYM.IMI$AUC))

##                                                 sensitivity         specificity 
##    "Optimal n LYM:"                "25"   "0.2181146025878" "0.864039408866995" 
##                                                             
##         "0.4609375" "0.674615384615385" "0.528903781539386"

print(c("Optimal %LYM:", PER.LYM.IMI$optimal_cutpoint, PER.LYM.IMI$sensitivity, PER.LYM.IMI$specificity, PER.LYM.IMI$ppv, PER.LYM.IMI$npv, PER.LYM.IMI$AUC))

##                                                 sensitivity         specificity 
##     "Optimal %LYM:"             "22.12" "0.484288354898336" "0.550738916256158" 
##                                                             
## "0.364902506963788" "0.667064439140811" "0.499978146654162"

print(c("Optimal n MAC:", exp(MAC.IMI$optimal_cutpoint), MAC.IMI$sensitivity, MAC.IMI$specificity, MAC.IMI$ppv, MAC.IMI$npv, MAC.IMI$AUC))

##                                                 sensitivity         specificity 
##    "Optimal n MAC:"                "32" "0.151571164510166" "0.905418719211823" 
##                                                             
## "0.460674157303371" "0.666908563134978" "0.517645666208354"

print(c("Optimal %MAC:", PER.MAC.IMI$optimal_cutpoint, PER.MAC.IMI$sensitivity, PER.MAC.IMI$specificity, PER.MAC.IMI$ppv, PER.MAC.IMI$npv, PER.MAC.IMI$AUC))

##                                                 sensitivity         specificity 
##     "Optimal %MAC:"             "26.92" "0.335731414868106" "0.713464696223317" 
##                                                             
## "0.286298568507157" "0.758289703315881" "0.523704189357873"

# Bootstrap confidence intervals
# SCC.IMI
SCC.IMI_optimal_cutpoint <- boot_ci(SCC.IMI, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
SCC.IMI_lower_ci <- subset(SCC.IMI_optimal_cutpoint, SCC.IMI_optimal_cutpoint$quantile == 0.025)$values
exp(SCC.IMI_lower_ci)

## [1] 86

SCC.IMI_upper_ci <- subset(SCC.IMI_optimal_cutpoint, SCC.IMI_optimal_cutpoint$quantile == 0.975)$values
exp(SCC.IMI_upper_ci)

## [1] 247

boot_ci(SCC.IMI, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.110
## 2    0.975  0.381

boot_ci(SCC.IMI, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.707
## 2    0.975  0.950

boot_ci(SCC.IMI, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.387
## 2    0.975  0.540

boot_ci(SCC.IMI, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.650
## 2    0.975  0.703

boot_ci(SCC.IMI, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.484
## 2    0.975  0.545

# NEU.IMI
NEU.IMI_optimal_cutpoint <- boot_ci(NEU.IMI, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
NEU.IMI_lower_ci <- subset(NEU.IMI_optimal_cutpoint, NEU.IMI_optimal_cutpoint$quantile == 0.025)$values
exp(NEU.IMI_lower_ci)

## [1] 34

NEU.IMI_upper_ci <- subset(NEU.IMI_optimal_cutpoint, NEU.IMI_optimal_cutpoint$quantile == 0.975)$values
exp(NEU.IMI_upper_ci)

## [1] 183

boot_ci(NEU.IMI, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.102
## 2    0.975  0.416

boot_ci(NEU.IMI, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.672
## 2    0.975  0.954

boot_ci(NEU.IMI, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.388
## 2    0.975  0.557

boot_ci(NEU.IMI, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.650
## 2    0.975  0.703

boot_ci(NEU.IMI, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.494
## 2    0.975  0.555

# PER.NEU.IMI
PER.NEU.IMI_optimal_cutpoint <- boot_ci(PER.NEU.IMI, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
PER.NEU.IMI_optimal_cutpoint

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025   44.9
## 2    0.975   72.1

PER.NEU.IMI_lower_ci <- subset(PER.NEU.IMI_optimal_cutpoint, PER.NEU.IMI_optimal_cutpoint$quantile == 0.025)$values
PER.NEU.IMI_upper_ci <- subset(PER.NEU.IMI_optimal_cutpoint, PER.NEU.IMI_optimal_cutpoint$quantile == 0.975)$values

boot_ci(PER.NEU.IMI, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025 0.0943
## 2    0.975 0.896

boot_ci(PER.NEU.IMI, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.161
## 2    0.975  0.945

boot_ci(PER.NEU.IMI, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.345
## 2    0.975  0.475

boot_ci(PER.NEU.IMI, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.647
## 2    0.975  0.743

boot_ci(PER.NEU.IMI, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.488
## 2    0.975  0.550

# LYM.IMI
LYM.IMI_optimal_cutpoint <- boot_ci(LYM.IMI, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
LYM.IMI_lower_ci <- subset(LYM.IMI_optimal_cutpoint, LYM.IMI_optimal_cutpoint$quantile == 0.025)$values
exp(LYM.IMI_lower_ci)

## [1] 12

LYM.IMI_upper_ci <- subset(LYM.IMI_optimal_cutpoint, LYM.IMI_optimal_cutpoint$quantile == 0.975)$values
exp(LYM.IMI_upper_ci)

## [1] 27

boot_ci(LYM.IMI, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.179
## 2    0.975  0.509

boot_ci(LYM.IMI, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.585
## 2    0.975  0.894

boot_ci(LYM.IMI, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.378
## 2    0.975  0.522

boot_ci(LYM.IMI, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.651
## 2    0.975  0.706

boot_ci(LYM.IMI, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.498
## 2    0.975  0.560

# PER.LYM.IMI
PER.LYM.IMI_optimal_cutpoint <- boot_ci(PER.LYM.IMI, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
PER.LYM.IMI_optimal_cutpoint

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025   18.7
## 2    0.975   39.0

PER.LYM.IMI_lower_ci <- subset(PER.LYM.IMI_optimal_cutpoint, PER.LYM.IMI_optimal_cutpoint$quantile == 0.025)$values
PER.LYM.IMI_upper_ci <- subset(PER.LYM.IMI_optimal_cutpoint, PER.LYM.IMI_optimal_cutpoint$quantile == 0.975)$values

boot_ci(PER.LYM.IMI, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025 0.0334
## 2    0.975 0.649

boot_ci(PER.LYM.IMI, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.394
## 2    0.975  0.979

boot_ci(PER.LYM.IMI, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.342
## 2    0.975  0.467

boot_ci(PER.LYM.IMI, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.641
## 2    0.975  0.704

boot_ci(PER.LYM.IMI, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.470
## 2    0.975  0.530

# MAC.IMI
MAC.IMI_optimal_cutpoint <- boot_ci(MAC.IMI, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
MAC.IMI_lower_ci <- subset(MAC.IMI_optimal_cutpoint, MAC.IMI_optimal_cutpoint$quantile == 0.025)$values
exp(MAC.IMI_lower_ci)

## [1] 10

MAC.IMI_upper_ci <- subset(MAC.IMI_optimal_cutpoint, MAC.IMI_optimal_cutpoint$quantile == 0.975)$values
exp(MAC.IMI_upper_ci)

## [1] 38

boot_ci(MAC.IMI, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.126
## 2    0.975  0.534

boot_ci(MAC.IMI, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.535
## 2    0.975  0.926

boot_ci(MAC.IMI, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.366
## 2    0.975  0.529

boot_ci(MAC.IMI, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.646
## 2    0.975  0.703

boot_ci(MAC.IMI, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.488
## 2    0.975  0.548

# PER.MAC.IMI
PER.MAC.IMI_optimal_cutpoint <- boot_ci(PER.MAC.IMI, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
PER.MAC.IMI_optimal_cutpoint

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025   7.79
## 2    0.975  34.6

PER.MAC.IMI_lower_ci <- subset(PER.MAC.IMI_optimal_cutpoint, PER.MAC.IMI_optimal_cutpoint$quantile == 0.025)$values
PER.MAC.IMI_upper_ci <- subset(PER.MAC.IMI_optimal_cutpoint, PER.MAC.IMI_optimal_cutpoint$quantile == 0.975)$values

boot_ci(PER.MAC.IMI, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.183
## 2    0.975  0.957

boot_ci(PER.MAC.IMI, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025 0.0938
## 2    0.975 0.876

boot_ci(PER.MAC.IMI, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.251
## 2    0.975  0.344

boot_ci(PER.MAC.IMI, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.744
## 2    0.975  0.874

boot_ci(PER.MAC.IMI, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.493
## 2    0.975  0.557

# Prevalence

# True prevalence

table(Pc.by.quarter.C5$IMI)

## 
##    0    1 
## 1020  543

prop.table(table(Pc.by.quarter.C5$IMI))

## 
##         0         1 
## 0.6525912 0.3474088

#Apparent prevalence

# ---- SCC ----

# optimal cut-off

optimal_scc_C5_IMI_quarter <- SCC.IMI$optimal_cutpoint

table(Pc.by.quarter.C5$test_result_n > exp(optimal_scc_C5_IMI_quarter))

## 
## FALSE  TRUE 
##  1394   248

prop.table(table(Pc.by.quarter.C5$test_result_n > exp(optimal_scc_C5_IMI_quarter)))

## 
##     FALSE      TRUE 
## 0.8489647 0.1510353

# confidence intervals

table(Pc.by.quarter.C5$test_result_n > exp(SCC.IMI_lower_ci))

## 
## FALSE  TRUE 
##  1128   514

prop.table(table(Pc.by.quarter.C5$test_result_n > exp(SCC.IMI_lower_ci)))

## 
##     FALSE      TRUE 
## 0.6869671 0.3130329

table(Pc.by.quarter.C5$test_result_n > exp(SCC.IMI_upper_ci))

## 
## FALSE  TRUE 
##  1524   118

prop.table(table(Pc.by.quarter.C5$test_result_n >exp(SCC.IMI_upper_ci)))

## 
##      FALSE       TRUE 
## 0.92813642 0.07186358

# ---- Neutrophils ----

# optimal cut-off

optimal_neu_C5_IMI_quarter <- NEU.IMI$optimal_cutpoint

table(Pc.by.quarter.C5$TOTAL01_N > exp(optimal_neu_C5_IMI_quarter))

## 
## FALSE  TRUE 
##  1310   337

prop.table(table(Pc.by.quarter.C5$TOTAL01_N > exp(optimal_neu_C5_IMI_quarter)))

## 
##     FALSE      TRUE 
## 0.7953855 0.2046145

# confidence intervals

table(Pc.by.quarter.C5$TOTAL01_N > exp(NEU.IMI_lower_ci))

## 
## FALSE  TRUE 
##  1071   576

prop.table(table(Pc.by.quarter.C5$TOTAL01_N > exp(NEU.IMI_lower_ci)))

## 
##     FALSE      TRUE 
## 0.6502732 0.3497268

table(Pc.by.quarter.C5$TOTAL01_N > exp(NEU.IMI_upper_ci))

## 
## FALSE  TRUE 
##  1552    95

prop.table(table(Pc.by.quarter.C5$TOTAL01_N >exp(NEU.IMI_upper_ci)))

## 
##      FALSE       TRUE 
## 0.94231937 0.05768063

# ---- Percent neu ----

# Optimal cut-off

optimal_per_neu_C5_IMI_quarter <- PER.NEU.IMI$optimal_cutpoint

table(Pc.by.quarter.C5$PERCENT01_N > optimal_per_neu_C5_IMI_quarter)

## 
## FALSE  TRUE 
##  1091   556

prop.table(table(Pc.by.quarter.C5$PERCENT01_N > optimal_per_neu_C5_IMI_quarter))

## 
##     FALSE      TRUE 
## 0.6624165 0.3375835

# confidence intervals

table(Pc.by.quarter.C5$PERCENT01_N > PER.NEU.IMI_lower_ci)

## 
## FALSE  TRUE 
##   228  1419

prop.table(table(Pc.by.quarter.C5$PERCENT01_N > PER.NEU.IMI_lower_ci))

## 
##     FALSE      TRUE 
## 0.1384335 0.8615665

table(Pc.by.quarter.C5$PERCENT01_N > PER.NEU.IMI_upper_ci)

## 
## FALSE  TRUE 
##  1543   104

prop.table(table(Pc.by.quarter.C5$PERCENT01_N >PER.NEU.IMI_upper_ci))

## 
##      FALSE       TRUE 
## 0.93685489 0.06314511

# ---- Lymphocytes ----

# optimal cut-off

optimal_lym_C5_IMI_quarter <-LYM.IMI$optimal_cutpoint

table(Pc.by.quarter.C5$TOTAL02_N > exp(optimal_lym_C5_IMI_quarter))

## 
## FALSE  TRUE 
##  1378   269

prop.table(table(Pc.by.quarter.C5$TOTAL02_N > exp(optimal_lym_C5_IMI_quarter)))

## 
##     FALSE      TRUE 
## 0.8366727 0.1633273

# confidence intervals

table(Pc.by.quarter.C5$TOTAL02_N > exp(LYM.IMI_lower_ci))

## 
## FALSE  TRUE 
##   995   652

prop.table(table(Pc.by.quarter.C5$TOTAL02_N > exp(LYM.IMI_lower_ci)))

## 
##     FALSE      TRUE 
## 0.6041287 0.3958713

table(Pc.by.quarter.C5$TOTAL02_N > exp(LYM.IMI_upper_ci))

## 
## FALSE  TRUE 
##  1422   225

prop.table(table(Pc.by.quarter.C5$TOTAL02_N >exp(LYM.IMI_upper_ci)))

## 
##    FALSE     TRUE 
## 0.863388 0.136612

# ---- Percent lym ----

# optimal cut-off

optimal_per_lym_C5_IMI_quarter <-PER.LYM.IMI$optimal_cutpoint

table(Pc.by.quarter.C5$PERCENT02_N > optimal_per_lym_C5_IMI_quarter)

## 
## FALSE  TRUE 
##   894   753

prop.table(table(Pc.by.quarter.C5$PERCENT02_N > optimal_per_lym_C5_IMI_quarter))

## 
##     FALSE      TRUE 
## 0.5428051 0.4571949

# confidence intervals

table(Pc.by.quarter.C5$PERCENT02_N > PER.LYM.IMI_lower_ci)

## 
## FALSE  TRUE 
##   624  1023

prop.table(table(Pc.by.quarter.C5$PERCENT02_N > PER.LYM.IMI_lower_ci))

## 
##     FALSE      TRUE 
## 0.3788707 0.6211293

table(Pc.by.quarter.C5$PERCENT02_N > PER.LYM.IMI_upper_ci)

## 
## FALSE  TRUE 
##  1601    46

prop.table(table(Pc.by.quarter.C5$PERCENT02_N >PER.LYM.IMI_upper_ci))

## 
##      FALSE       TRUE 
## 0.97207043 0.02792957

# ---- Macrophages ----

# optimal cut-off

optimal_mac_C5_IMI_quarter <- MAC.IMI$optimal_cutpoint

table(Pc.by.quarter.C5$TOTAL03_N > exp(optimal_mac_C5_IMI_quarter))

## 
## FALSE  TRUE 
##  1464   183

prop.table(table(Pc.by.quarter.C5$TOTAL03_N > exp(optimal_mac_C5_IMI_quarter)))

## 
##     FALSE      TRUE 
## 0.8888889 0.1111111

# confidence intervals

table(Pc.by.quarter.C5$TOTAL03_N > exp(MAC.IMI_lower_ci))

## 
## FALSE  TRUE 
##   927   720

prop.table(table(Pc.by.quarter.C5$TOTAL03_N > exp(MAC.IMI_lower_ci)))

## 
##     FALSE      TRUE 
## 0.5628415 0.4371585

table(Pc.by.quarter.C5$TOTAL03_N > exp(MAC.IMI_upper_ci))

## 
## FALSE  TRUE 
##  1488   159

prop.table(table(Pc.by.quarter.C5$TOTAL03_N >exp(MAC.IMI_upper_ci)))

## 
##      FALSE       TRUE 
## 0.90346084 0.09653916

# Percent MAC

# optimal cut-off

optimal_per_mac_C5_IMI_quarter <- PER.MAC.IMI$optimal_cutpoint

table(Pc.by.quarter.C5$PERCENT03_N > optimal_per_mac_C5_IMI_quarter)

## 
## FALSE  TRUE 
##  1154   493

prop.table(table(Pc.by.quarter.C5$PERCENT03_N > optimal_per_mac_C5_IMI_quarter))

## 
##     FALSE      TRUE 
## 0.7006679 0.2993321

# confidence intervals

table(Pc.by.quarter.C5$PERCENT03_N > PER.MAC.IMI_lower_ci)

## 
## FALSE  TRUE 
##   132  1515

prop.table(table(Pc.by.quarter.C5$PERCENT03_N > PER.MAC.IMI_lower_ci))

## 
##      FALSE       TRUE 
## 0.08014572 0.91985428

table(Pc.by.quarter.C5$PERCENT03_N > PER.MAC.IMI_upper_ci)

## 
## FALSE  TRUE 
##  1419   228

prop.table(table(Pc.by.quarter.C5$PERCENT03_N >PER.MAC.IMI_upper_ci))

## 
##     FALSE      TRUE 
## 0.8615665 0.1384335

Quarter high SCC

# Set a seed for reproducibility

set.seed(123)

# Qscout subclinical mastitis reference

# n NEU SCM quarter
NEU.delaval.SCM <- cutpointr(Pc.by.quarter.C5, log.Neu, delaval.SCM, 
                                 method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(NEU.delaval.SCM)

## Method: maximize_metric 
## Predictor: log.Neu 
## Outcome: delaval.SCM 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##    AUC    n n_pos n_neg
##  0.936 1639   144  1495
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity   ppv    npv  tp
##            4.4308        1.8015 0.9506      0.8403      0.9612 0.676 0.9842 121
##  fn fp   tn
##  23 58 1437
## 
## Predictor summary: 
##     Data     Min.       5%  1st Qu.   Median     Mean  3rd Qu.      95%
##  Overall 1.098612 2.185446 2.772589 3.258097 3.420589 3.806662 5.293305
##        0 1.098612 2.079442 2.772589 3.178054 3.199462 3.650574 4.356709
##        1 1.386294 3.306556 4.764247 5.379632 5.716317 6.676715 8.923461
##      Max.        SD NAs
##  10.17477 1.1011778   0
##   8.40268 0.7124932   0
##  10.17477 1.6636133   0
## 
## Bootstrap summary: 
##           Variable Min.   5% 1st Qu. Median Mean 3rd Qu.  95% Max.   SD NAs
##   optimal_cutpoint 4.25 4.25    4.43   4.43 4.43    4.44 4.48 4.74 0.06   0
##              AUC_b 0.88 0.91    0.93   0.94 0.94    0.95 0.96 0.97 0.01   0
##            AUC_oob 0.88 0.91    0.92   0.94 0.94    0.95 0.96 0.99 0.02   0
##    sum_sens_spec_b 1.69 1.75    1.78   1.80 1.80    1.82 1.85 1.91 0.03   0
##  sum_sens_spec_oob 1.60 1.71    1.76   1.79 1.79    1.82 1.86 1.93 0.04   0
##              acc_b 0.91 0.93    0.95   0.95 0.95    0.96 0.96 0.98 0.01   0
##            acc_oob 0.90 0.92    0.94   0.95 0.95    0.95 0.96 0.98 0.01   0
##      sensitivity_b 0.73 0.79    0.82   0.84 0.84    0.86 0.89 0.94 0.03   0
##    sensitivity_oob 0.64 0.75    0.80   0.83 0.83    0.86 0.90 0.98 0.05   0
##      specificity_b 0.92 0.94    0.96   0.96 0.96    0.97 0.97 0.99 0.01   0
##    specificity_oob 0.91 0.93    0.96   0.96 0.96    0.97 0.97 0.99 0.01   0
##     cohens_kappa_b 0.58 0.65    0.70   0.72 0.72    0.75 0.78 0.85 0.04   0
##   cohens_kappa_oob 0.50 0.62    0.68   0.71 0.71    0.74 0.78 0.87 0.05   0
##              ppv_b 0.49 0.57    0.65   0.68 0.67    0.70 0.75 0.90 0.05   0
##            ppv_oob 0.41 0.55    0.63   0.67 0.67    0.71 0.77 0.92 0.06   0
##              npv_b 0.97 0.98    0.98   0.98 0.98    0.99 0.99 0.99 0.00   0
##            npv_oob 0.96 0.98    0.98   0.98 0.98    0.99 0.99 1.00 0.00   0

# ROC plot
plot(NEU.delaval.SCM)

# Youden index plot
plot_metric(NEU.delaval.SCM)

# %NEU SCM quarter
PER.NEU.delaval.SCM <- cutpointr(Pc.by.quarter.C5, PERCENT01_N, delaval.SCM, 
                                 method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(PER.NEU.delaval.SCM)

## Method: maximize_metric 
## Predictor: PERCENT01_N 
## Outcome: delaval.SCM 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC    n n_pos n_neg
##  0.5201 1639   144  1495
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv tp
##             60.79        1.1287 0.6632      0.4444      0.6843 0.1194 0.9275 64
##  fn  fp   tn
##  80 472 1023
## 
## Predictor summary: 
##     Data  Min.     5% 1st Qu. Median     Mean 3rd Qu.     95%  Max.       SD
##  Overall 13.79 38.090  49.740 56.320 56.24826   63.27 73.6800 89.55 10.70931
##        0 13.79 38.668  50.000 56.250 56.25045   62.91 73.6090 89.55 10.45925
##        1 21.89 32.215  46.665 58.425 56.22542   66.01 74.6815 78.90 13.06880
##  NAs
##    0
##    0
##    0
## 
## Bootstrap summary: 
##           Variable  Min.    5% 1st Qu. Median  Mean 3rd Qu.   95%  Max.   SD
##   optimal_cutpoint 57.44 59.65   60.79  61.18 62.17   63.67 65.50 72.37 2.25
##              AUC_b  0.42  0.47    0.50   0.52  0.52    0.54  0.57  0.63 0.03
##            AUC_oob  0.37  0.46    0.49   0.52  0.52    0.54  0.58  0.65 0.04
##    sum_sens_spec_b  1.00  1.08    1.12   1.15  1.15    1.17  1.21  1.28 0.04
##  sum_sens_spec_oob  0.84  1.00    1.05   1.09  1.09    1.12  1.18  1.28 0.05
##              acc_b  0.54  0.62    0.65   0.69  0.70    0.74  0.78  0.88 0.05
##            acc_oob  0.54  0.61    0.65   0.68  0.69    0.73  0.78  0.87 0.06
##      sensitivity_b  0.10  0.28    0.37   0.43  0.42    0.48  0.54  0.64 0.08
##    sensitivity_oob  0.02  0.21    0.31   0.38  0.37    0.43  0.50  0.62 0.09
##      specificity_b  0.53  0.63    0.67   0.71  0.72    0.77  0.82  0.95 0.07
##    specificity_oob  0.55  0.62    0.67   0.71  0.72    0.77  0.82  0.96 0.07
##     cohens_kappa_b  0.00  0.04    0.06   0.07  0.07    0.09  0.11  0.16 0.02
##   cohens_kappa_oob -0.08  0.00    0.03   0.04  0.04    0.06  0.09  0.14 0.03
##              ppv_b  0.09  0.11    0.12   0.13  0.13    0.14  0.16  0.25 0.02
##            ppv_oob  0.02  0.08    0.10   0.11  0.11    0.13  0.15  0.19 0.02
##              npv_b  0.90  0.92    0.92   0.93  0.93    0.93  0.94  0.95 0.01
##            npv_oob  0.89  0.91    0.92   0.92  0.92    0.93  0.94  0.96 0.01
##  NAs
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0

# ROC plot
plot(PER.NEU.delaval.SCM)

# Youden index plot
plot_metric(PER.NEU.delaval.SCM)

# LYM SCM quarter
LYM.delaval.SCM <- cutpointr(Pc.by.quarter.C5, log.Lym, delaval.SCM, method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(LYM.delaval.SCM)

## Method: maximize_metric 
## Predictor: log.Lym 
## Outcome: delaval.SCM 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC    n n_pos n_neg
##  0.9314 1639   144  1495
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv
##            3.2189         1.782 0.9042       0.875       0.907 0.4755 0.9869
##   tp fn  fp   tn
##  126 18 139 1356
## 
## Predictor summary: 
##     Data     Min.       5%  1st Qu.   Median     Mean  3rd Qu.      95%
##  Overall 0.000000 1.098612 1.791759 2.302585 2.421411 2.890372 4.317488
##        0 0.000000 1.098612 1.609438 2.197225 2.219420 2.708050 3.433987
##        1 1.609438 2.213029 3.637586 4.330733 4.518462 5.395708 7.259279
##      Max.        SD NAs
##  8.165932 1.0720204   0
##  7.006695 0.7766316   0
##  8.165932 1.4168435   0
## 
## Bootstrap summary: 
##           Variable Min.   5% 1st Qu. Median Mean 3rd Qu.  95% Max.   SD NAs
##   optimal_cutpoint 3.09 3.22    3.22   3.22 3.28    3.33 3.40 3.74 0.09   0
##              AUC_b 0.88 0.91    0.92   0.93 0.93    0.94 0.95 0.98 0.01   0
##            AUC_oob 0.87 0.90    0.92   0.93 0.93    0.95 0.96 0.99 0.02   0
##    sum_sens_spec_b 1.67 1.74    1.77   1.79 1.79    1.81 1.83 1.88 0.03   0
##  sum_sens_spec_oob 1.57 1.69    1.74   1.77 1.76    1.80 1.83 1.90 0.04   0
##              acc_b 0.86 0.90    0.90   0.91 0.92    0.93 0.94 0.97 0.02   0
##            acc_oob 0.84 0.89    0.90   0.91 0.91    0.92 0.94 0.96 0.02   0
##      sensitivity_b 0.74 0.81    0.84   0.87 0.87    0.89 0.92 0.97 0.03   0
##    sensitivity_oob 0.61 0.75    0.81   0.85 0.85    0.88 0.93 0.98 0.05   0
##      specificity_b 0.86 0.90    0.91   0.92 0.92    0.94 0.95 0.98 0.02   0
##    specificity_oob 0.85 0.89    0.90   0.92 0.92    0.93 0.95 0.98 0.02   0
##     cohens_kappa_b 0.46 0.53    0.56   0.59 0.60    0.64 0.70 0.80 0.05   0
##   cohens_kappa_oob 0.37 0.50    0.55   0.58 0.58    0.62 0.67 0.74 0.05   0
##              ppv_b 0.38 0.44    0.47   0.50 0.52    0.56 0.63 0.80 0.06   0
##            ppv_oob 0.30 0.41    0.46   0.50 0.51    0.54 0.62 0.77 0.07   0
##              npv_b 0.97 0.98    0.98   0.99 0.99    0.99 0.99 1.00 0.00   0
##            npv_oob 0.96 0.97    0.98   0.98 0.98    0.99 0.99 1.00 0.01   0

# ROC plot
plot(LYM.delaval.SCM)

# Youden index plot
plot_metric(LYM.delaval.SCM)

# %LYM SCM quarter
PER.LYM.delaval.SCM <- cutpointr(Pc.by.quarter.C5, PERCENT02_N, delaval.SCM, 
                                 method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 0

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(PER.LYM.delaval.SCM)

## Method: maximize_metric 
## Predictor: PERCENT02_N 
## Outcome: delaval.SCM 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC    n n_pos n_neg
##  0.6575 1639  1495   144
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv
##             21.54        1.2785 0.5308      0.5077      0.7708 0.9583 0.1311
##   tp  fn fp  tn
##  759 736 33 111
## 
## Predictor summary: 
##     Data Min.     5% 1st Qu. Median     Mean 3rd Qu.    95%  Max.       SD NAs
##  Overall 0.00  9.800 16.1450 21.210 21.87254  27.075 36.567 61.54 8.148360   0
##        0 0.00 10.000 16.4750 21.670 22.25621  27.530 37.033 61.54 8.190777   0
##        1 1.52  9.013 13.4675 17.505 17.88938  21.440 29.937 35.38 6.488743   0
## 
## Bootstrap summary: 
##           Variable  Min.    5% 1st Qu. Median  Mean 3rd Qu.   95%  Max.   SD
##   optimal_cutpoint 14.55 20.00   20.95  21.54 21.37   22.03 22.45 27.14 1.01
##              AUC_b  0.58  0.62    0.64   0.66  0.66    0.67  0.69  0.72 0.02
##            AUC_oob  0.55  0.61    0.64   0.66  0.66    0.68  0.71  0.75 0.03
##    sum_sens_spec_b  1.17  1.23    1.26   1.29  1.29    1.31  1.34  1.42 0.04
##  sum_sens_spec_oob  1.00  1.15    1.21   1.25  1.25    1.29  1.34  1.45 0.06
##              acc_b  0.32  0.49    0.51   0.53  0.54    0.56  0.60  0.80 0.04
##            acc_oob  0.31  0.48    0.51   0.53  0.53    0.55  0.60  0.77 0.04
##      sensitivity_b  0.26  0.46    0.49   0.51  0.52    0.54  0.59  0.83 0.05
##    sensitivity_oob  0.25  0.45    0.48   0.51  0.51    0.53  0.59  0.83 0.05
##      specificity_b  0.40  0.68    0.74   0.78  0.77    0.81  0.84  0.93 0.06
##    specificity_oob  0.24  0.59    0.70   0.75  0.74    0.79  0.84  0.95 0.09
##     cohens_kappa_b  0.04  0.07    0.08   0.09  0.09    0.10  0.12  0.20 0.02
##   cohens_kappa_oob  0.00  0.05    0.07   0.08  0.08    0.09  0.11  0.15 0.02
##              ppv_b  0.93  0.95    0.95   0.96  0.96    0.96  0.97  0.98 0.01
##            ppv_oob  0.90  0.93    0.95   0.95  0.95    0.96  0.97  0.99 0.01
##              npv_b  0.09  0.11    0.13   0.13  0.13    0.14  0.15  0.23 0.01
##            npv_oob  0.08  0.10    0.12   0.13  0.13    0.14  0.15  0.19 0.02
##  NAs
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0

# ROC plot
plot(PER.LYM.delaval.SCM)

# Youden index plot
plot_metric(PER.LYM.delaval.SCM)

# MAC SCM quarter
MAC.delaval.SCM <- cutpointr(Pc.by.quarter.C5, log.Mac, delaval.SCM, method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(MAC.delaval.SCM)

## Method: maximize_metric 
## Predictor: log.Mac 
## Outcome: delaval.SCM 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC    n n_pos n_neg
##  0.9497 1639   144  1495
## 
##  optimal_cutpoint sum_sens_spec   acc sensitivity specificity    ppv    npv  tp
##            3.0445        1.7784 0.878      0.9028      0.8756 0.4114 0.9894 130
##  fn  fp   tn
##  14 186 1309
## 
## Predictor summary: 
##     Data     Min.        5%  1st Qu.   Median     Mean  3rd Qu.      95%
##  Overall 0.000000 0.6931472 1.609438 2.197225 2.345712 2.772589 4.443821
##        0 0.000000 0.6931472 1.609438 2.079442 2.108938 2.639057 3.443512
##        1 1.791759 2.5649494 3.555348 4.599842 4.803890 5.860774 7.942045
##      Max.        SD NAs
##  9.127176 1.1999069   0
##  7.458186 0.8311637   0
##  9.127176 1.6135293   0
## 
## Bootstrap summary: 
##           Variable Min.   5% 1st Qu. Median Mean 3rd Qu.  95% Max.   SD NAs
##   optimal_cutpoint 3.04 3.04    3.04   3.09 3.12    3.18 3.26 3.89 0.09   0
##              AUC_b 0.90 0.93    0.94   0.95 0.95    0.96 0.97 0.98 0.01   0
##            AUC_oob 0.91 0.93    0.94   0.95 0.95    0.96 0.97 0.99 0.01   0
##    sum_sens_spec_b 1.67 1.74    1.77   1.78 1.78    1.80 1.83 1.89 0.03   0
##  sum_sens_spec_oob 1.52 1.69    1.74   1.77 1.76    1.79 1.82 1.89 0.04   0
##              acc_b 0.86 0.87    0.88   0.89 0.89    0.90 0.92 0.97 0.02   0
##            acc_oob 0.84 0.86    0.88   0.89 0.89    0.90 0.91 0.95 0.02   0
##      sensitivity_b 0.73 0.84    0.87   0.89 0.89    0.91 0.94 0.98 0.03   0
##    sensitivity_oob 0.54 0.78    0.84   0.88 0.87    0.91 0.94 1.00 0.05   0
##      specificity_b 0.85 0.87    0.88   0.89 0.89    0.91 0.92 0.99 0.02   0
##    specificity_oob 0.84 0.86    0.88   0.89 0.89    0.91 0.92 0.99 0.02   0
##     cohens_kappa_b 0.41 0.47    0.51   0.54 0.54    0.57 0.61 0.81 0.05   0
##   cohens_kappa_oob 0.33 0.45    0.49   0.52 0.52    0.56 0.60 0.73 0.05   0
##              ppv_b 0.33 0.38    0.41   0.44 0.45    0.48 0.52 0.83 0.05   0
##            ppv_oob 0.26 0.36    0.40   0.43 0.44    0.47 0.52 0.81 0.05   0
##              npv_b 0.98 0.98    0.99   0.99 0.99    0.99 0.99 1.00 0.00   0
##            npv_oob 0.96 0.98    0.98   0.99 0.99    0.99 0.99 1.00 0.01   0

# ROC plot
plot(MAC.delaval.SCM)

# Youden index plot
plot_metric(MAC.delaval.SCM)

# %MAC SCM quarter
PER.MAC.delaval.SCM<- cutpointr(Pc.by.quarter.C5, PERCENT03_N, delaval.SCM, method = maximize_metric, metric = sum_sens_spec,na.rm=TRUE, boot_runs = 2000) %>% add_metric(list(ppv, npv)) 

## Assuming the positive class is 1

## Assuming the positive class has higher x values

## Running bootstrap...

# Summary
summary(PER.MAC.delaval.SCM)

## Method: maximize_metric 
## Predictor: PERCENT03_N 
## Outcome: delaval.SCM 
## Direction: >= 
## Nr. of bootstraps: 2000 
## 
##     AUC    n n_pos n_neg
##  0.5875 1639   144  1495
## 
##  optimal_cutpoint sum_sens_spec    acc sensitivity specificity    ppv    npv tp
##             26.39        1.1878 0.6827      0.4861      0.7017 0.1357 0.9341 70
##  fn  fp   tn
##  74 446 1049
## 
## Predictor summary: 
##     Data Min.    5% 1st Qu. Median     Mean 3rd Qu.     95%  Max.       SD NAs
##  Overall 0.00 6.450 13.0400  20.00 21.87892  28.800 43.5590 71.43 11.60963   0
##        0 0.00 6.441 13.0400  19.77 21.49295  28.310 41.9520 71.43 11.29463   0
##        1 4.99 7.109 13.7875  25.17 25.88604  35.085 50.7025 69.06 13.90600   0
## 
## Bootstrap summary: 
##           Variable  Min.    5% 1st Qu. Median  Mean 3rd Qu.   95%  Max.   SD
##   optimal_cutpoint 17.90 22.84   23.53  26.39 26.71   29.58 31.70 41.24 3.28
##              AUC_b  0.47  0.54    0.57   0.59  0.59    0.61  0.63  0.67 0.03
##            AUC_oob  0.45  0.53    0.56   0.59  0.59    0.61  0.65  0.71 0.04
##    sum_sens_spec_b  1.06  1.14    1.18   1.20  1.20    1.23  1.27  1.33 0.04
##  sum_sens_spec_oob  0.95  1.06    1.11   1.15  1.15    1.18  1.24  1.32 0.06
##              acc_b  0.47  0.60    0.63   0.69  0.69    0.75  0.79  0.89 0.07
##            acc_oob  0.45  0.59    0.62   0.67  0.68    0.74  0.79  0.87 0.07
##      sensitivity_b  0.14  0.34    0.43   0.51  0.50    0.57  0.63  0.73 0.09
##    sensitivity_oob  0.10  0.27    0.37   0.45  0.44    0.52  0.60  0.69 0.10
##      specificity_b  0.44  0.60    0.63   0.70  0.71    0.78  0.83  0.95 0.08
##    specificity_oob  0.43  0.59    0.63   0.70  0.70    0.78  0.83  0.95 0.08
##     cohens_kappa_b  0.03  0.06    0.08   0.10  0.10    0.12  0.15  0.22 0.03
##   cohens_kappa_oob -0.02  0.02    0.05   0.07  0.07    0.09  0.12  0.20 0.03
##              ppv_b  0.09  0.12    0.13   0.14  0.14    0.16  0.19  0.27 0.02
##            ppv_oob  0.06  0.09    0.11   0.13  0.13    0.14  0.17  0.28 0.02
##              npv_b  0.91  0.92    0.93   0.94  0.94    0.94  0.95  0.96 0.01
##            npv_oob  0.88  0.91    0.92   0.93  0.93    0.94  0.95  0.96 0.01
##  NAs
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0
##    0

# ROC plot
plot(PER.MAC.delaval.SCM)

# Youden index plot
plot_metric(PER.MAC.delaval.SCM)

# Summary of results when using SCM as reference

print(c("Optimal n NEU:", exp(NEU.delaval.SCM$optimal_cutpoint), NEU.delaval.SCM$sensitivity, NEU.delaval.SCM$specificity, NEU.delaval.SCM$ppv, NEU.delaval.SCM$npv, NEU.delaval.SCM$AUC))

##                                                 sensitivity         specificity 
##    "Optimal n NEU:"                "84" "0.840277777777778" "0.961204013377926" 
##                                                             
## "0.675977653631285" "0.984246575342466" "0.936022853957637"

print(c("Optimal %NEU:", PER.NEU.delaval.SCM$optimal_cutpoint, PER.NEU.delaval.SCM$sensitivity, PER.NEU.delaval.SCM$specificity, PER.NEU.delaval.SCM$ppv, PER.NEU.delaval.SCM$npv, PER.NEU.delaval.SCM$AUC))

##                                                 sensitivity         specificity 
##     "Optimal %NEU:"             "60.79" "0.444444444444444"  "0.68428093645485" 
##                                                             
## "0.119402985074627" "0.927470534904805" "0.520080824972129"

print(c("Optimal n LYM:", exp(LYM.delaval.SCM$optimal_cutpoint), LYM.delaval.SCM$sensitivity, LYM.delaval.SCM$specificity, LYM.delaval.SCM$ppv, LYM.delaval.SCM$npv, LYM.delaval.SCM$AUC))

##                                                 sensitivity         specificity 
##    "Optimal n LYM:"                "25"             "0.875" "0.907023411371237" 
##                                                             
## "0.475471698113208" "0.986899563318777" "0.931400966183575"

print(c("Optimal %LYM:", PER.LYM.delaval.SCM$optimal_cutpoint, PER.LYM.delaval.SCM$sensitivity, PER.LYM.delaval.SCM$specificity, PER.LYM.delaval.SCM$ppv, PER.LYM.delaval.SCM$npv, PER.LYM.delaval.SCM$AUC))

##                                                 sensitivity         specificity 
##     "Optimal %LYM:"             "21.54" "0.507692307692308" "0.770833333333333" 
##                                                             
## "0.958333333333333" "0.131050767414404" "0.657541341508733"

print(c("Optimal n MAC:", exp(MAC.delaval.SCM$optimal_cutpoint), MAC.delaval.SCM$sensitivity, MAC.delaval.SCM$specificity, MAC.delaval.SCM$ppv, MAC.delaval.SCM$npv, MAC.delaval.SCM$AUC))

##                                                 sensitivity         specificity 
##    "Optimal n MAC:"                "21" "0.902777777777778" "0.875585284280936" 
##                                                             
## "0.411392405063291" "0.989417989417989" "0.949716648086213"

print(c("Optimal %MAC:", PER.MAC.delaval.SCM$optimal_cutpoint, PER.MAC.delaval.SCM$sensitivity, PER.MAC.delaval.SCM$specificity, PER.MAC.delaval.SCM$ppv, PER.MAC.delaval.SCM$npv, PER.MAC.delaval.SCM$AUC))

##                                                 sensitivity         specificity 
##     "Optimal %MAC:"             "26.39" "0.486111111111111" "0.701672240802676" 
##                                                             
## "0.135658914728682" "0.934105075690116" "0.587541806020067"

# CI using bootstrapping
# NEU.delaval.SCM
NEU.delaval.SCM_optimal_cutpoint <- boot_ci(NEU.delaval.SCM, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
NEU.delaval.SCM_lower_ci <- subset(NEU.delaval.SCM_optimal_cutpoint, NEU.delaval.SCM_optimal_cutpoint$quantile == 0.025)$values
exp(NEU.delaval.SCM_lower_ci)

## [1] 70

NEU.delaval.SCM_upper_ci <- subset(NEU.delaval.SCM_optimal_cutpoint, NEU.delaval.SCM_optimal_cutpoint$quantile == 0.975)$values
exp(NEU.delaval.SCM_upper_ci)

## [1] 99

boot_ci(NEU.delaval.SCM, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.775
## 2    0.975  0.899

boot_ci(NEU.delaval.SCM, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.933
## 2    0.975  0.976

boot_ci(NEU.delaval.SCM, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.547
## 2    0.975  0.774

boot_ci(NEU.delaval.SCM, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.977
## 2    0.975  0.990

boot_ci(NEU.delaval.SCM, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.908
## 2    0.975  0.960

# PER.NEU.delaval.SCM
PER.NEU.delaval.SCM_optimal_cutpoint <- boot_ci(PER.NEU.delaval.SCM, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
PER.NEU.delaval.SCM_optimal_cutpoint

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025   59.2
## 2    0.975   67.3

PER.NEU.delaval.SCM_lower_ci <- subset(PER.NEU.delaval.SCM_optimal_cutpoint, PER.NEU.delaval.SCM_optimal_cutpoint$quantile == 0.025)$values
PER.NEU.delaval.SCM_upper_ci <- subset(PER.NEU.delaval.SCM_optimal_cutpoint, PER.NEU.delaval.SCM_optimal_cutpoint$quantile == 0.975)$values

boot_ci(PER.NEU.delaval.SCM, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.253
## 2    0.975  0.557

boot_ci(PER.NEU.delaval.SCM, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.619
## 2    0.975  0.863

boot_ci(PER.NEU.delaval.SCM, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.101
## 2    0.975  0.167

boot_ci(PER.NEU.delaval.SCM, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.914
## 2    0.975  0.944

boot_ci(PER.NEU.delaval.SCM, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.464
## 2    0.975  0.578

# LYM.delaval.SCM
LYM.delaval.SCM_optimal_cutpoint <- boot_ci(LYM.delaval.SCM, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
LYM.delaval.SCM_lower_ci <- subset(LYM.delaval.SCM_optimal_cutpoint, LYM.delaval.SCM_optimal_cutpoint$quantile == 0.025)$values
exp(LYM.delaval.SCM_lower_ci)

## [1] 25

LYM.delaval.SCM_upper_ci <- subset(LYM.delaval.SCM_optimal_cutpoint, LYM.delaval.SCM_optimal_cutpoint$quantile == 0.975)$values
exp(LYM.delaval.SCM_upper_ci)

## [1] 35

boot_ci(LYM.delaval.SCM, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.8  
## 2    0.975  0.923

boot_ci(LYM.delaval.SCM, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.895
## 2    0.975  0.963

boot_ci(LYM.delaval.SCM, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.424
## 2    0.975  0.674

boot_ci(LYM.delaval.SCM, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.979
## 2    0.975  0.992

boot_ci(LYM.delaval.SCM, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.901
## 2    0.975  0.958

# PER.LYM.delaval.SCM
PER.LYM.delaval.SCM_optimal_cutpoint <- boot_ci(PER.LYM.delaval.SCM, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
PER.LYM.delaval.SCM_optimal_cutpoint

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025   18.2
## 2    0.975   22.4

PER.LYM.delaval.SCM_lower_ci <- subset(PER.LYM.delaval.SCM_optimal_cutpoint, PER.LYM.delaval.SCM_optimal_cutpoint$quantile == 0.025)$values
PER.LYM.delaval.SCM_upper_ci <- subset(PER.LYM.delaval.SCM_optimal_cutpoint, PER.LYM.delaval.SCM_optimal_cutpoint$quantile == 0.975)$values

boot_ci(PER.LYM.delaval.SCM, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.452
## 2    0.975  0.681

boot_ci(PER.LYM.delaval.SCM, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.610
## 2    0.975  0.857

boot_ci(PER.LYM.delaval.SCM, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.943
## 2    0.975  0.973

boot_ci(PER.LYM.delaval.SCM, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.110
## 2    0.975  0.160

boot_ci(PER.LYM.delaval.SCM, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.611
## 2    0.975  0.698

# MAC.delaval.SCM
MAC.delaval.SCM_optimal_cutpoint <- boot_ci(MAC.delaval.SCM, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
MAC.delaval.SCM_lower_ci <- subset(MAC.delaval.SCM_optimal_cutpoint, MAC.delaval.SCM_optimal_cutpoint$quantile == 0.025)$values
exp(MAC.delaval.SCM_lower_ci)

## [1] 21

MAC.delaval.SCM_upper_ci <- subset(MAC.delaval.SCM_optimal_cutpoint, MAC.delaval.SCM_optimal_cutpoint$quantile == 0.975)$values
exp(MAC.delaval.SCM_upper_ci)

## [1] 26

boot_ci(MAC.delaval.SCM, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.832
## 2    0.975  0.942

boot_ci(MAC.delaval.SCM, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.863
## 2    0.975  0.927

boot_ci(MAC.delaval.SCM, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.369
## 2    0.975  0.545

boot_ci(MAC.delaval.SCM, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.982
## 2    0.975  0.994

boot_ci(MAC.delaval.SCM, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.930
## 2    0.975  0.968

# PER.MAC.delaval.SCM
PER.MAC.delaval.SCM_optimal_cutpoint <- boot_ci(PER.MAC.delaval.SCM, optimal_cutpoint, in_bag = TRUE, alpha = 0.05)
PER.MAC.delaval.SCM_optimal_cutpoint

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025   22.8
## 2    0.975   32.1

PER.MAC.delaval.SCM_lower_ci <- subset(PER.MAC.delaval.SCM_optimal_cutpoint, PER.MAC.delaval.SCM_optimal_cutpoint$quantile == 0.025)$values
PER.MAC.delaval.SCM_upper_ci <- subset(PER.MAC.delaval.SCM_optimal_cutpoint, PER.MAC.delaval.SCM_optimal_cutpoint$quantile == 0.975)$values

boot_ci(PER.MAC.delaval.SCM, sensitivity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.312
## 2    0.975  0.646

boot_ci(PER.MAC.delaval.SCM, specificity, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.592
## 2    0.975  0.844

boot_ci(PER.MAC.delaval.SCM, ppv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.111
## 2    0.975  0.195

boot_ci(PER.MAC.delaval.SCM, npv, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.921
## 2    0.975  0.951

boot_ci(PER.MAC.delaval.SCM, AUC, in_bag = TRUE, alpha = 0.05)

## # A tibble: 2 × 2
##   quantile values
##      <dbl>  <dbl>
## 1    0.025  0.535
## 2    0.975  0.642

# Prevalence

# True prevalence

table(Pc.by.quarter.C5$delaval.SCM)

## 
##    0    1 
## 1497  145

prop.table(table(Pc.by.quarter.C5$delaval.SCM))

## 
##          0          1 
## 0.91169306 0.08830694

#Apparent prevalence

# ---- Neutrophils ----

# optimal cut-off

optimal_neu_C5_delaval.SCM_quarter <- NEU.delaval.SCM$optimal_cutpoint

table(Pc.by.quarter.C5$TOTAL01_N > exp(optimal_neu_C5_delaval.SCM_quarter))

## 
## FALSE  TRUE 
##  1464   183

prop.table(table(Pc.by.quarter.C5$TOTAL01_N > exp(optimal_neu_C5_delaval.SCM_quarter)))

## 
##     FALSE      TRUE 
## 0.8888889 0.1111111

# confidence intervals

table(Pc.by.quarter.C5$TOTAL01_N > exp(NEU.delaval.SCM_lower_ci))

## 
## FALSE  TRUE 
##  1424   223

prop.table(table(Pc.by.quarter.C5$TOTAL01_N > exp(NEU.delaval.SCM_lower_ci)))

## 
##     FALSE      TRUE 
## 0.8646023 0.1353977

table(Pc.by.quarter.C5$TOTAL01_N > exp(NEU.delaval.SCM_upper_ci))

## 
## FALSE  TRUE 
##  1492   155

prop.table(table(Pc.by.quarter.C5$TOTAL01_N >exp(NEU.delaval.SCM_upper_ci)))

## 
##     FALSE      TRUE 
## 0.9058895 0.0941105

# ---- Percent neu ----

# Optimal cut-off

optimal_per_neu_C5_delaval.SCM_quarter <- PER.NEU.delaval.SCM$optimal_cutpoint

table(Pc.by.quarter.C5$PERCENT01_N > optimal_per_neu_C5_delaval.SCM_quarter)

## 
## FALSE  TRUE 
##  1111   536

prop.table(table(Pc.by.quarter.C5$PERCENT01_N > optimal_per_neu_C5_delaval.SCM_quarter))

## 
##     FALSE      TRUE 
## 0.6745598 0.3254402

# confidence intervals

table(Pc.by.quarter.C5$PERCENT01_N > PER.NEU.delaval.SCM_lower_ci)

## 
## FALSE  TRUE 
##  1007   640

prop.table(table(Pc.by.quarter.C5$PERCENT01_N > PER.NEU.delaval.SCM_lower_ci))

## 
##     FALSE      TRUE 
## 0.6114147 0.3885853

table(Pc.by.quarter.C5$PERCENT01_N > PER.NEU.delaval.SCM_upper_ci)

## 
## FALSE  TRUE 
##  1407   240

prop.table(table(Pc.by.quarter.C5$PERCENT01_N >PER.NEU.delaval.SCM_upper_ci))

## 
##     FALSE      TRUE 
## 0.8542805 0.1457195

# ---- Lymphocytes ----

# optimal cut-off

optimal_lym_C5_delaval.SCM_quarter <-LYM.delaval.SCM$optimal_cutpoint

table(Pc.by.quarter.C5$TOTAL02_N > exp(optimal_lym_C5_delaval.SCM_quarter))

## 
## FALSE  TRUE 
##  1378   269

prop.table(table(Pc.by.quarter.C5$TOTAL02_N > exp(optimal_lym_C5_delaval.SCM_quarter)))

## 
##     FALSE      TRUE 
## 0.8366727 0.1633273

# confidence intervals

table(Pc.by.quarter.C5$TOTAL02_N > exp(LYM.delaval.SCM_lower_ci))

## 
## FALSE  TRUE 
##  1378   269

prop.table(table(Pc.by.quarter.C5$TOTAL02_N > exp(LYM.delaval.SCM_lower_ci)))

## 
##     FALSE      TRUE 
## 0.8366727 0.1633273

table(Pc.by.quarter.C5$TOTAL02_N > exp(LYM.delaval.SCM_upper_ci))

## 
## FALSE  TRUE 
##  1476   171

prop.table(table(Pc.by.quarter.C5$TOTAL02_N >exp(LYM.delaval.SCM_upper_ci)))

## 
##     FALSE      TRUE 
## 0.8961749 0.1038251

# ---- Percent lym ----

# optimal cut-off

optimal_per_lym_C5_delaval.SCM_quarter <-PER.LYM.delaval.SCM$optimal_cutpoint

table(Pc.by.quarter.C5$PERCENT02_N > optimal_per_lym_C5_delaval.SCM_quarter)

## 
## FALSE  TRUE 
##   849   798

prop.table(table(Pc.by.quarter.C5$PERCENT02_N > optimal_per_lym_C5_delaval.SCM_quarter))

## 
##     FALSE      TRUE 
## 0.5154827 0.4845173

# confidence intervals

table(Pc.by.quarter.C5$PERCENT02_N > PER.LYM.delaval.SCM_lower_ci)

## 
## FALSE  TRUE 
##   590  1057

prop.table(table(Pc.by.quarter.C5$PERCENT02_N > PER.LYM.delaval.SCM_lower_ci))

## 
##     FALSE      TRUE 
## 0.3582271 0.6417729

table(Pc.by.quarter.C5$PERCENT02_N > PER.LYM.delaval.SCM_upper_ci)

## 
## FALSE  TRUE 
##   928   719

prop.table(table(Pc.by.quarter.C5$PERCENT02_N >PER.LYM.delaval.SCM_upper_ci))

## 
##     FALSE      TRUE 
## 0.5634487 0.4365513

# ---- Macrophages ----

# optimal cut-off

optimal_mac_C5_delaval.SCM_quarter <- MAC.delaval.SCM$optimal_cutpoint

table(Pc.by.quarter.C5$TOTAL03_N > exp(optimal_mac_C5_delaval.SCM_quarter))

## 
## FALSE  TRUE 
##  1346   301

prop.table(table(Pc.by.quarter.C5$TOTAL03_N > exp(optimal_mac_C5_delaval.SCM_quarter)))

## 
##     FALSE      TRUE 
## 0.8172435 0.1827565

# confidence intervals

table(Pc.by.quarter.C5$TOTAL03_N > exp(MAC.delaval.SCM_lower_ci))

## 
## FALSE  TRUE 
##  1346   301

prop.table(table(Pc.by.quarter.C5$TOTAL03_N > exp(MAC.delaval.SCM_lower_ci)))

## 
##     FALSE      TRUE 
## 0.8172435 0.1827565

table(Pc.by.quarter.C5$TOTAL03_N > exp(MAC.delaval.SCM_upper_ci))

## 
## FALSE  TRUE 
##  1417   230

prop.table(table(Pc.by.quarter.C5$TOTAL03_N >exp(MAC.delaval.SCM_upper_ci)))

## 
##     FALSE      TRUE 
## 0.8603522 0.1396478

# Percent MAC

# optimal cut-off

optimal_per_mac_C5_delaval.SCM_quarter <- PER.MAC.delaval.SCM$optimal_cutpoint

table(Pc.by.quarter.C5$PERCENT03_N > optimal_per_mac_C5_delaval.SCM_quarter)

## 
## FALSE  TRUE 
##  1132   515

prop.table(table(Pc.by.quarter.C5$PERCENT03_N > optimal_per_mac_C5_delaval.SCM_quarter))

## 
##     FALSE      TRUE 
## 0.6873103 0.3126897

# confidence intervals

table(Pc.by.quarter.C5$PERCENT03_N > PER.MAC.delaval.SCM_lower_ci)

## 
## FALSE  TRUE 
##   974   673

prop.table(table(Pc.by.quarter.C5$PERCENT03_N > PER.MAC.delaval.SCM_lower_ci))

## 
##     FALSE      TRUE 
## 0.5913783 0.4086217

table(Pc.by.quarter.C5$PERCENT03_N > PER.MAC.delaval.SCM_upper_ci)

## 
## FALSE  TRUE 
##  1358   289

prop.table(table(Pc.by.quarter.C5$PERCENT03_N > PER.MAC.delaval.SCM_upper_ci))

## 
##     FALSE      TRUE 
## 0.8245294 0.1754706

Plots

# ---- Plots ----

# Make sure outcomes are recognized as a factor


Doff.by.quarter.2$IMI <- as.factor(Doff.by.quarter.2$IMI)
Doff.by.quarter.2$delaval.SCM <- as.factor(Doff.by.quarter.2$delaval.SCM)


Pc.by.quarter.C3$IMI <- as.factor(Pc.by.quarter.C3$IMI)
Pc.by.quarter.C3$delaval.SCM <- as.factor(Pc.by.quarter.C3$delaval.SCM)


Pc.by.quarter.C5$IMI <- as.factor(Pc.by.quarter.C5$IMI)
Pc.by.quarter.C5$delaval.SCM <- as.factor(Pc.by.quarter.C5$delaval.SCM)



Doff.by.cow.3$IMI.cow <- as.factor(Doff.by.cow.3$IMI.cow)
Doff.by.cow.3$delaval.SCM.CO.COW <- as.factor(Doff.by.cow.3$delaval.SCM.CO.COW)


Pc.by.cow.C3$IMI.cow <- as.factor(Pc.by.cow.C3$IMI.cow)
Pc.by.cow.C3$delaval.SCM.CO.COW <- as.factor(Pc.by.cow.C3$delaval.SCM.CO.COW)


Pc.by.cow.C5$IMI.cow <- as.factor(Pc.by.cow.C5$IMI.cow)
Pc.by.cow.C5$delaval.SCM.CO.COW <- as.factor(Pc.by.cow.C5$delaval.SCM.CO.COW)


# ---- Detection of IMI quarter level ----

# Neu 

theme_set(theme_bw())

# Dry-off

IMI.1_quarter_doff <- Doff.by.quarter.2 %>%
filter(!is.na(log.Neu) & !is.na(IMI)) %>%
ggplot(aes(x=log.Neu, group=IMI, fill=IMI)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,8) +
  labs(x="log-NEU/mL",title="") +
  geom_vline(xintercept = optimal_neu_doff_IMI_quarter, linetype="dotted", size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 

## Warning: Using `size` aesthetic for lines was deprecated in ggplot2 3.4.0.
## ℹ Please use `linewidth` instead.
## This warning is displayed once every 8 hours.
## Call `lifecycle::last_lifecycle_warnings()` to see where this warning was
## generated.

# Lym

IMI.2_quarter_doff <-Doff.by.quarter.2 %>%
  filter(!is.na(log.Lym) & !is.na(IMI)) %>%
  ggplot(aes(x=log.Lym, group=IMI, fill=IMI)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,8) +
  labs(x="log-LYM/mL",title="") +
  geom_vline(xintercept = optimal_lym_doff_IMI_quarter, linetype="dotted", 
              size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 


# Mac 

IMI.3_quarter_doff <- Doff.by.quarter.2 %>%
  filter(!is.na(log.Mac) & !is.na(IMI)) %>%
  ggplot(aes(x=log.Mac, group=IMI, fill=IMI)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,8) +
  labs(x="log-MAC/mL",title="") +
  geom_vline(xintercept = optimal_mac_doff_IMI_quarter, linetype="dotted", 
              size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 


# % Neu

IMI.4_quarter_doff <- Doff.by.quarter.2 %>%
  filter(!is.na(PERCENT01_N) & !is.na(IMI)) %>%
  ggplot(aes(x=PERCENT01_N, group=IMI, fill=IMI)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,100) +
  labs(x="% NEU",title="") +
  geom_vline(xintercept = optimal_per_neu_doff_IMI_quarter, linetype="dotted", 
              size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 


# % Lym

IMI.5_quarter_doff <- Doff.by.quarter.2 %>%
  filter(!is.na(PERCENT02_N) & !is.na(IMI)) %>%
  ggplot(aes(x=PERCENT02_N, group=IMI, fill=IMI)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,100) +
  labs(x="% LYM",title="") +
  geom_vline(xintercept = optimal_per_lym_doff_IMI_quarter, linetype="dotted", 
              size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 


# % Mac

IMI.6_quarter_doff <- Doff.by.quarter.2 %>%
  filter(!is.na(PERCENT03_N) & !is.na(IMI)) %>%
  ggplot(aes(x=PERCENT03_N, group=IMI, fill=IMI)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,100) +
  labs(x="% MAC",title="") +
  geom_vline(xintercept = optimal_per_mac_doff_IMI_quarter, linetype="dotted", 
              size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 

# Multipanel figure

IMI_quarter_doff <- ggarrange(IMI.1_quarter_doff + rremove("ylab"), IMI.2_quarter_doff + rremove("ylab"), IMI.3_quarter_doff + rremove("ylab"),
                      IMI.4_quarter_doff + rremove("ylab"), IMI.5_quarter_doff + rremove("ylab"), IMI.6_quarter_doff + rremove("ylab"),
                           ncol = 3, nrow = 2, common.legend = TRUE,align = "h",labels = c("A1", "B1", "C1","D1","E1","F1","G1"))

## Warning: Removed 18 rows containing non-finite outside the scale range
## (`stat_density()`).

## Warning: Removed 18 rows containing non-finite outside the scale range
## (`stat_density()`).

## Warning: Removed 1 row containing non-finite outside the scale range
## (`stat_density()`).

## Warning: Removed 5 rows containing non-finite outside the scale range
## (`stat_density()`).

IMI_quarter_doff_annotated <- annotate_figure(IMI_quarter_doff,
                          top = text_grob("Dry-off", color = "black",face="bold",size=20,
                          hjust = 0, x = 0,y=-0.75))

IMI_quarter_doff_annotated

ggsave(plot = last_plot(),"./figures/DSCC.IMI_quarter_doff.png",width = 20, height = 10, units = "cm") 


# ---- C3 -----

# Neu

IMI.1_quarter_C3 <- Pc.by.quarter.C3 %>%
  filter(!is.na(log.Neu) & !is.na(IMI)) %>%
  ggplot(aes(x=log.Neu, group=IMI, fill=IMI)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,8) +
  labs(x="log-NEU/mL",title="") +
  theme(legend.position = "none") +
  geom_vline(xintercept = optimal_neu_C3_IMI_quarter, linetype="dotted", size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 

# Lym

IMI.2_quarter_C3 <-Pc.by.quarter.C3 %>%
  filter(!is.na(log.Lym) & !is.na(IMI)) %>%
  ggplot(aes(x=log.Lym, group=IMI, fill=IMI)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,8) +
  labs(x="log-LYM/mL",title="") +
  theme(legend.position = "none") +
  geom_vline(xintercept = optimal_lym_C3_IMI_quarter, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 


# Mac 

IMI.3_quarter_C3 <- Pc.by.quarter.C3 %>%
  filter(!is.na(log.Mac) & !is.na(IMI)) %>%
  ggplot(aes(x=log.Mac, group=IMI, fill=IMI)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,8) +
  labs(x="log-MAC/mL",title="") +
  theme(legend.position = "none") +
  geom_vline(xintercept = optimal_mac_C3_IMI_quarter, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 


# % Neu

IMI.4_quarter_C3 <- Pc.by.quarter.C3 %>%
  filter(!is.na(PERCENT01_N) & !is.na(IMI)) %>%
  ggplot(aes(x=PERCENT01_N, group=IMI, fill=IMI)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,100) +
  labs(x="% NEU",title="") +
  theme(legend.position = "none") +
  geom_vline(xintercept = optimal_per_neu_C3_IMI_quarter, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 


# % Lym

IMI.5_quarter_C3 <- Pc.by.quarter.C3 %>%
  filter(!is.na(PERCENT02_N) & !is.na(IMI)) %>%
  ggplot(aes(x=PERCENT02_N, group=IMI, fill=IMI)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,100) +
  labs(x="% LYM",title="") +
  theme(legend.position = "none") +
  geom_vline(xintercept = optimal_per_lym_C3_IMI_quarter, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 


# % Mac

IMI.6_quarter_C3 <- Pc.by.quarter.C3 %>%
  filter(!is.na(PERCENT03_N) & !is.na(IMI)) %>%
  ggplot(aes(x=PERCENT03_N, group=IMI, fill=IMI)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,100) +
  labs(x="% MAC",title="") +
  theme(legend.position = "none") +
  geom_vline(xintercept = optimal_per_mac_C3_IMI_quarter, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 

# Multipanel figure

IMI_quarter_C3 <- ggarrange(IMI.1_quarter_C3 + rremove("ylab"), IMI.2_quarter_C3 + rremove("ylab"), IMI.3_quarter_C3 + rremove("ylab"),
                              IMI.4_quarter_C3 + rremove("ylab"), IMI.5_quarter_C3 + rremove("ylab"), IMI.6_quarter_C3 + rremove("ylab"),
                              ncol = 3, nrow = 2,align = "h",labels = c("A2", "B2", "C2","D2","E2","F2","G2"))

## Warning: Removed 11 rows containing non-finite outside the scale range
## (`stat_density()`).

## Warning: Removed 3 rows containing non-finite outside the scale range
## (`stat_density()`).

IMI_quarter_C3_annotated <- annotate_figure(IMI_quarter_C3,
                                              top = text_grob("Postpartum: 3 DIM", color = "black",face="bold",size=20,
                                                              hjust = 0, x = 0))

IMI_quarter_C3_annotated

ggsave(plot = last_plot(),"./figures/DSCC.IMI_quarter_C3.png",width = 20, height = 10, units = "cm") 


# ---- c5 ----

# Neu

IMI.1_quarter_C5 <- Pc.by.quarter.C5 %>%
  filter(!is.na(log.Neu) & !is.na(IMI)) %>%
  ggplot(aes(x=log.Neu, group=IMI, fill=IMI)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,8) +
  labs(x="log-NEU/mL",title="") +
  theme(legend.position = "none") +
  geom_vline(xintercept = optimal_neu_C5_IMI_quarter, linetype="dotted", size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 

# Lym

IMI.2_quarter_C5 <-Pc.by.quarter.C5 %>%
  filter(!is.na(log.Lym) & !is.na(IMI)) %>%
  ggplot(aes(x=log.Lym, group=IMI, fill=IMI)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,8) +
  labs(x="log-LYM/mL",title="") +
  theme(legend.position = "none") +
  geom_vline(xintercept = optimal_lym_C5_IMI_quarter, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 


# Mac 

IMI.3_quarter_C5 <- Pc.by.quarter.C5 %>%
  filter(!is.na(log.Mac) & !is.na(IMI)) %>%
  ggplot(aes(x=log.Mac, group=IMI, fill=IMI)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,8) +
  labs(x="log-MAC/mL",title="") +
  theme(legend.position = "none") +
  geom_vline(xintercept = optimal_mac_C5_IMI_quarter, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 


# % Neu

IMI.4_quarter_C5 <- Pc.by.quarter.C5 %>%
  filter(!is.na(PERCENT01_N) & !is.na(IMI)) %>%
  ggplot(aes(x=PERCENT01_N, group=IMI, fill=IMI)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,100) +
  labs(x="% NEU",title="") +
  theme(legend.position = "none") +
  geom_vline(xintercept = optimal_per_neu_C5_IMI_quarter, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 


# % Lym

IMI.5_quarter_C5 <- Pc.by.quarter.C5 %>%
  filter(!is.na(PERCENT02_N) & !is.na(IMI)) %>%
  ggplot(aes(x=PERCENT02_N, group=IMI, fill=IMI)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,100) +
  labs(x="% LYM",title="") +
  theme(legend.position = "none") +
  geom_vline(xintercept = optimal_per_lym_C5_IMI_quarter, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 


# % Mac

IMI.6_quarter_C5 <- Pc.by.quarter.C5 %>%
  filter(!is.na(PERCENT03_N) & !is.na(IMI)) %>%
  ggplot(aes(x=PERCENT03_N, group=IMI, fill=IMI)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,100) +
  labs(x="% MAC",title="") +
  theme(legend.position = "none") +
  geom_vline(xintercept = optimal_per_mac_C5_IMI_quarter, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 

# Multipanel figure

IMI_quarter_C5 <- ggarrange(IMI.1_quarter_C5 + rremove("ylab"), IMI.2_quarter_C5 + rremove("ylab"), IMI.3_quarter_C5 + rremove("ylab"),
                            IMI.4_quarter_C5 + rremove("ylab"), IMI.5_quarter_C5 + rremove("ylab"), IMI.6_quarter_C5 + rremove("ylab"),
                            ncol = 3, nrow = 2,align = "h",labels = c("A3", "B3", "C3","D3","E3","F3","G3"))

## Warning: Removed 18 rows containing non-finite outside the scale range
## (`stat_density()`).

## Warning: Removed 1 row containing non-finite outside the scale range
## (`stat_density()`).

## Warning: Removed 6 rows containing non-finite outside the scale range
## (`stat_density()`).

IMI_quarter_C5_annotated <- annotate_figure(IMI_quarter_C5,
                                            top = text_grob("Postpartum: 5 DIM", color = "black",face="bold",size=20,
                                                            hjust = 0, x = 0))

IMI_quarter_C5_annotated

ggsave(plot = last_plot(),"./figures/DSCC.IMI_quarter_C5.png",width = 20, height = 10, units = "cm") 



DSCC.IMI_all <- ggarrange(IMI_quarter_doff_annotated + rremove("ylab"), IMI_quarter_C3_annotated + rremove("ylab"), IMI_quarter_C5_annotated + rremove("ylab"),
                              ncol = 1, nrow = 3, common.legend = TRUE,align = "h",widths = 4)

DSCC.IMI_all

ggsave(plot = last_plot(),"./figures/DSCC.IMI_quarter_all.png",width = 20, height = 30, units = "cm")

# ---- Subclinical mastitis quarter level ----

# Neu 

SCM.1_quarter_doff <- Doff.by.quarter.2 %>%
  filter(!is.na(log.Neu) & !is.na(delaval.SCM)) %>%
  ggplot(aes(x=log.Neu, group=delaval.SCM, fill=delaval.SCM)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,8) +
  labs(x="log-NEU/mL",title="",fill="High SCC") +
  geom_vline(xintercept = optimal_neu_doff_delaval.SCM_quarter, linetype="dotted", size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 

# Lym

SCM.2_quarter_doff <-Doff.by.quarter.2 %>%
  filter(!is.na(log.Lym) & !is.na(delaval.SCM)) %>%
  ggplot(aes(x=log.Lym, group=delaval.SCM, fill=delaval.SCM)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,8) +
  labs(x="log-LYM/mL",title="",fill="High SCC") +
  geom_vline(xintercept = optimal_lym_doff_delaval.SCM_quarter, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 


# Mac 

SCM.3_quarter_doff <- Doff.by.quarter.2 %>%
  filter(!is.na(log.Mac) & !is.na(delaval.SCM)) %>%
  ggplot(aes(x=log.Mac, group=delaval.SCM, fill=delaval.SCM)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,8) +
  labs(x="log-MAC/mL",title="",fill="High SCC") +
  geom_vline(xintercept = optimal_mac_doff_delaval.SCM_quarter, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 


# % Neu

SCM.4_quarter_doff <- Doff.by.quarter.2 %>%
  filter(!is.na(PERCENT01_N) & !is.na(delaval.SCM)) %>%
  ggplot(aes(x=PERCENT01_N, group=delaval.SCM, fill=delaval.SCM)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,100) +
  labs(x="% NEU",title="",fill="High SCC") +
  geom_vline(xintercept = optimal_per_neu_doff_delaval.SCM_quarter, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 


# % Lym

SCM.5_quarter_doff <- Doff.by.quarter.2 %>%
  filter(!is.na(PERCENT02_N) & !is.na(delaval.SCM)) %>%
  ggplot(aes(x=PERCENT02_N, group=delaval.SCM, fill=delaval.SCM)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,100) +
  labs(x="% LYM",title="",fill="High SCC") +
  geom_vline(xintercept = optimal_per_lym_doff_delaval.SCM_quarter, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 


# % Mac

SCM.6_quarter_doff <- Doff.by.quarter.2 %>%
  filter(!is.na(PERCENT03_N) & !is.na(delaval.SCM)) %>%
  ggplot(aes(x=PERCENT03_N, group=delaval.SCM, fill=delaval.SCM)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,100) +
  labs(x="% MAC",title="",fill="High SCC") +
  geom_vline(xintercept = optimal_per_mac_doff_delaval.SCM_quarter, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 

# Multipanel figure

delaval.SCM_quarter_doff <- ggarrange(SCM.1_quarter_doff + rremove("ylab"), SCM.2_quarter_doff + rremove("ylab"), SCM.3_quarter_doff + rremove("ylab"),
                              SCM.4_quarter_doff + rremove("ylab"), SCM.5_quarter_doff + rremove("ylab"), SCM.6_quarter_doff + rremove("ylab"),
                              ncol = 3, nrow = 2, common.legend = TRUE,align = "h",labels = c("A1", "B1", "C1","D1","E1","F1","G1"))

## Warning: Removed 18 rows containing non-finite outside the scale range
## (`stat_density()`).

## Warning: Removed 18 rows containing non-finite outside the scale range
## (`stat_density()`).

## Warning: Removed 1 row containing non-finite outside the scale range
## (`stat_density()`).

## Warning: Removed 5 rows containing non-finite outside the scale range
## (`stat_density()`).

delaval.SCM_quarter_doff_annotated <- annotate_figure(delaval.SCM_quarter_doff,
                                              top = text_grob("Dry-off", color = "black",face="bold",size=20,
                                                              hjust = 0, x = 0,y=-0.75))

delaval.SCM_quarter_doff_annotated

ggsave(plot = last_plot(),"./figures/DSCC.delaval.SCM_quarter_doff.png",width = 20, height = 10, units = "cm") 


# ---- C3 -----

# Neu

SCM.1_quarter_C3 <- Pc.by.quarter.C3 %>%
  filter(!is.na(log.Neu) & !is.na(delaval.SCM)) %>%
  ggplot(aes(x=log.Neu, group=delaval.SCM, fill=delaval.SCM)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,8) +
  labs(x="log-NEU/mL",title="",fill="High SCC") +
  theme(legend.position = "none") +
  geom_vline(xintercept = optimal_neu_C3_delaval.SCM_quarter, linetype="dotted", size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 

# Lym

SCM.2_quarter_C3 <-Pc.by.quarter.C3 %>%
  filter(!is.na(log.Lym) & !is.na(delaval.SCM)) %>%
  ggplot(aes(x=log.Lym, group=delaval.SCM, fill=delaval.SCM)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,8) +
  labs(x="log-LYM/mL",title="",fill="High SCC") +
  theme(legend.position = "none") +
  geom_vline(xintercept = optimal_lym_C3_delaval.SCM_quarter, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 


# Mac 

SCM.3_quarter_C3 <- Pc.by.quarter.C3 %>%
  filter(!is.na(log.Mac) & !is.na(delaval.SCM)) %>%
  ggplot(aes(x=log.Mac, group=delaval.SCM, fill=delaval.SCM)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,8) +
  labs(x="log-MAC/mL",title="",fill="High SCC") +
  theme(legend.position = "none") +
  geom_vline(xintercept = optimal_mac_C3_delaval.SCM_quarter, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 


# % Neu

SCM.4_quarter_C3 <- Pc.by.quarter.C3 %>%
  filter(!is.na(PERCENT01_N) & !is.na(delaval.SCM)) %>%
  ggplot(aes(x=PERCENT01_N, group=delaval.SCM, fill=delaval.SCM)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,100) +
  labs(x="% NEU",title="",fill="High SCC") +
  theme(legend.position = "none") +
  geom_vline(xintercept = optimal_per_neu_C3_delaval.SCM_quarter, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 


# % Lym

SCM.5_quarter_C3 <- Pc.by.quarter.C3 %>%
  filter(!is.na(PERCENT02_N) & !is.na(delaval.SCM)) %>%
  ggplot(aes(x=PERCENT02_N, group=delaval.SCM, fill=delaval.SCM)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,100) +
  labs(x="% LYM",title="",fill="High SCC") +
  theme(legend.position = "none") +
  geom_vline(xintercept = optimal_per_lym_C3_delaval.SCM_quarter, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 


# % Mac

SCM.6_quarter_C3 <- Pc.by.quarter.C3 %>%
  filter(!is.na(PERCENT03_N) & !is.na(delaval.SCM)) %>%
  ggplot(aes(x=PERCENT03_N, group=delaval.SCM, fill=delaval.SCM)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,100) +
  labs(x="% MAC",title="",fill="High SCC") +
  theme(legend.position = "none") +
  geom_vline(xintercept = optimal_per_mac_C3_delaval.SCM_quarter, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 

# Multipanel figure

delaval.SCM_quarter_C3 <- ggarrange(SCM.1_quarter_C3 + rremove("ylab"), SCM.2_quarter_C3 + rremove("ylab"), SCM.3_quarter_C3 + rremove("ylab"),
                            SCM.4_quarter_C3 + rremove("ylab"), SCM.5_quarter_C3 + rremove("ylab"), SCM.6_quarter_C3 + rremove("ylab"),
                            ncol = 3, nrow = 2,align = "h",labels = c("A2", "B2", "C2","D2","E2","F2","G2"))

## Warning: Removed 13 rows containing non-finite outside the scale range
## (`stat_density()`).

## Warning: Removed 3 rows containing non-finite outside the scale range
## (`stat_density()`).

delaval.SCM_quarter_C3_annotated <- annotate_figure(delaval.SCM_quarter_C3,
                                            top = text_grob("Postpartum: 3 DIM", color = "black",face="bold",size=20,
                                                            hjust = 0, x = 0))

delaval.SCM_quarter_C3_annotated

ggsave(plot = last_plot(),"./figures/DSCC.delaval.SCM_quarter_C3.png",width = 20, height = 10, units = "cm") 


# ---- c5 ----

# Neu

SCM.1_quarter_C5 <- Pc.by.quarter.C5 %>%
  filter(!is.na(log.Neu) & !is.na(delaval.SCM)) %>%
  ggplot(aes(x=log.Neu, group=delaval.SCM, fill=delaval.SCM)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,8) +
  labs(x="log-NEU/mL",title="",fill="High SCC") +
  theme(legend.position = "none") +
  geom_vline(xintercept = optimal_neu_C5_delaval.SCM_quarter, linetype="dotted", size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 

# Lym

SCM.2_quarter_C5 <-Pc.by.quarter.C5 %>%
  filter(!is.na(log.Lym) & !is.na(delaval.SCM)) %>%
  ggplot(aes(x=log.Lym, group=delaval.SCM, fill=delaval.SCM)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,8) +
  labs(x="log-LYM/mL",title="",fill="High SCC") +
  theme(legend.position = "none") +
  geom_vline(xintercept = optimal_lym_C5_delaval.SCM_quarter, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 


# Mac 

SCM.3_quarter_C5 <- Pc.by.quarter.C5 %>%
  filter(!is.na(log.Mac) & !is.na(delaval.SCM)) %>%
  ggplot(aes(x=log.Mac, group=delaval.SCM, fill=delaval.SCM)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,8) +
  labs(x="log-MAC/mL",title="",fill="High SCC") +
  theme(legend.position = "none") +
  geom_vline(xintercept = optimal_mac_C5_delaval.SCM_quarter, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 


# % Neu

SCM.4_quarter_C5 <- Pc.by.quarter.C5 %>%
  filter(!is.na(PERCENT01_N) & !is.na(delaval.SCM)) %>%
  ggplot(aes(x=PERCENT01_N, group=delaval.SCM, fill=delaval.SCM)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,100) +
  labs(x="% NEU",title="",fill="High SCC") +
  theme(legend.position = "none") +
  geom_vline(xintercept = optimal_per_neu_C5_delaval.SCM_quarter, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 


# % Lym

SCM.5_quarter_C5 <- Pc.by.quarter.C5 %>%
  filter(!is.na(PERCENT02_N) & !is.na(delaval.SCM)) %>%
  ggplot(aes(x=PERCENT02_N, group=delaval.SCM, fill=delaval.SCM)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,100) +
  labs(x="% LYM",title="",fill="High SCC") +
  theme(legend.position = "none") +
  geom_vline(xintercept = optimal_per_lym_C5_delaval.SCM_quarter, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 


# % Mac

SCM.6_quarter_C5 <- Pc.by.quarter.C5 %>%
  filter(!is.na(PERCENT03_N) & !is.na(delaval.SCM)) %>%
  ggplot(aes(x=PERCENT03_N, group=delaval.SCM, fill=delaval.SCM)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,100) +
  labs(x="% MAC",title="",fill="High SCC") +
  theme(legend.position = "none") +
  geom_vline(xintercept = optimal_per_mac_C5_delaval.SCM_quarter, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 

# Multipanel figure

delaval.SCM_quarter_C5 <- ggarrange(SCM.1_quarter_C5 + rremove("ylab"), SCM.2_quarter_C5 + rremove("ylab"), SCM.3_quarter_C5 + rremove("ylab"),
                            SCM.4_quarter_C5 + rremove("ylab"), SCM.5_quarter_C5 + rremove("ylab"), SCM.6_quarter_C5 + rremove("ylab"),
                            ncol = 3, nrow = 2,align = "h",labels = c("A3", "B3", "C3","D3","E3","F3","G3"))

## Warning: Removed 19 rows containing non-finite outside the scale range
## (`stat_density()`).

## Warning: Removed 1 row containing non-finite outside the scale range
## (`stat_density()`).

## Warning: Removed 7 rows containing non-finite outside the scale range
## (`stat_density()`).

delaval.SCM_quarter_C5_annotated <- annotate_figure(delaval.SCM_quarter_C5,
                                            top = text_grob("Postpartum: 5 DIM", color = "black",face="bold",size=20,
                                                            hjust = 0, x = 0))

delaval.SCM_quarter_C5_annotated

ggsave(plot = last_plot(),"./figures/DSCC.delaval.SCM_quarter_C5.png",width = 20, height = 10, units = "cm") 



DSCC.SCM_all <- ggarrange(delaval.SCM_quarter_doff_annotated + rremove("ylab"), delaval.SCM_quarter_C3_annotated + rremove("ylab"), delaval.SCM_quarter_C5_annotated + rremove("ylab"),
                          ncol = 1, nrow = 3, common.legend = TRUE,align = "h",widths = 4)

DSCC.SCM_all

ggsave(plot = last_plot(),"./figures/DSCC.delaval.SCM_quarter_all.png",width = 20, height = 30, units = "cm")


# ----- IMI cow level -----

# Neu 

IMI.1_cow_doff <- Doff.by.cow.3 %>%
  filter(!is.na(log.TOTAL.NEU.CO.COW) & !is.na(IMI.cow)) %>%
  ggplot(aes(x=log.TOTAL.NEU.CO.COW, group=IMI.cow, fill=IMI.cow)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,8) +
  labs(x="log-NEU/mL",title="",fille="IMI") +
  geom_vline(xintercept = optimal_neu_doff_IMI_cow, linetype="dotted", size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 

# Lym

IMI.2_cow_doff <-Doff.by.cow.3 %>%
  filter(!is.na(log.TOTAL.LYM.CO.COW) & !is.na(IMI.cow)) %>%
  ggplot(aes(x=log.TOTAL.LYM.CO.COW, group=IMI.cow, fill=IMI.cow)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,8) +
  labs(x="log-LYM/mL",title="",fille="IMI") +
  geom_vline(xintercept = optimal_lym_doff_IMI_cow, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 


# Mac 

IMI.3_cow_doff <- Doff.by.cow.3 %>%
  filter(!is.na(log.TOTAL.MAC.CO.COW) & !is.na(IMI.cow)) %>%
  ggplot(aes(x=log.TOTAL.MAC.CO.COW, group=IMI.cow, fill=IMI.cow)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,8) +
  labs(x="log-MAC/mL",title="",fille="IMI") +
  geom_vline(xintercept = optimal_mac_doff_IMI_cow, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 


# % Neu

IMI.4_cow_doff <- Doff.by.cow.3 %>%
  filter(!is.na(PERCENT.NEU.CO.COW) & !is.na(IMI.cow)) %>%
  ggplot(aes(x=PERCENT.NEU.CO.COW, group=IMI.cow, fill=IMI.cow)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,100) +
  labs(x="% NEU",title="",fille="IMI") +
  geom_vline(xintercept = optimal_per_neu_doff_IMI_cow, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 


# % Lym

IMI.5_cow_doff <- Doff.by.cow.3 %>%
  filter(!is.na(PERCENT.LYM.CO.COW) & !is.na(IMI.cow)) %>%
  ggplot(aes(x=PERCENT.LYM.CO.COW, group=IMI.cow, fill=IMI.cow)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,100) +
  labs(x="% LYM",title="",fille="IMI") +
  geom_vline(xintercept = optimal_per_lym_doff_IMI_cow, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 


# % Mac

IMI.6_cow_doff <- Doff.by.cow.3 %>%
  filter(!is.na(PERCENT.MAC.CO.COW) & !is.na(IMI.cow)) %>%
  ggplot(aes(x=PERCENT.MAC.CO.COW, group=IMI.cow, fill=IMI.cow)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,100) +
  labs(x="% MAC",title="",fille="IMI") +
  geom_vline(xintercept = optimal_per_mac_doff_IMI_cow, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 

# Multipanel figure

IMI_cow_doff <- ggarrange(IMI.1_cow_doff + rremove("ylab"), IMI.2_cow_doff + rremove("ylab"), IMI.3_cow_doff + rremove("ylab"),
                              IMI.4_cow_doff + rremove("ylab"), IMI.5_cow_doff + rremove("ylab"), IMI.6_cow_doff + rremove("ylab"),
                              ncol = 3, nrow = 2, common.legend = TRUE,align = "h",labels = c("A1", "B1", "C1","D1","E1","F1","G1"))

## Warning: Removed 5 rows containing non-finite outside the scale range
## (`stat_density()`).

## Warning: Removed 5 rows containing non-finite outside the scale range
## (`stat_density()`).

## Warning: Removed 1 row containing non-finite outside the scale range
## (`stat_density()`).

IMI_cow_doff_annotated <- annotate_figure(IMI_cow_doff,
                                              top = text_grob("Dry-off", color = "black",face="bold",size=20,
                                                              hjust = 0, x = 0,y=-0.75))

IMI_cow_doff_annotated

ggsave(plot = last_plot(),"./figures/DSCC.IMI_cow_doff.png",width = 20, height = 10, units = "cm") 


# ---- C3 -----

# Neu

IMI.1_cow_C3 <- Pc.by.cow.C3 %>%
  filter(!is.na(log.TOTAL.NEU.CO.COW) & !is.na(IMI.cow)) %>%
  ggplot(aes(x=log.TOTAL.NEU.CO.COW, group=IMI.cow, fill=IMI.cow)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,8) +
  labs(x="log-NEU/mL",title="",fille="IMI") +
  theme(legend.position = "none") +
  geom_vline(xintercept = optimal_neu_C3_IMI_cow, linetype="dotted", size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 

# Lym

IMI.2_cow_C3 <-Pc.by.cow.C3 %>%
  filter(!is.na(log.TOTAL.LYM.CO.COW) & !is.na(IMI.cow)) %>%
  ggplot(aes(x=log.TOTAL.LYM.CO.COW, group=IMI.cow, fill=IMI.cow)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,8) +
  labs(x="log-LYM/mL",title="",fille="IMI") +
  theme(legend.position = "none") +
  geom_vline(xintercept = optimal_lym_C3_IMI_cow, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 


# Mac 

IMI.3_cow_C3 <- Pc.by.cow.C3 %>%
  filter(!is.na(log.TOTAL.MAC.CO.COW) & !is.na(IMI.cow)) %>%
  ggplot(aes(x=log.TOTAL.MAC.CO.COW, group=IMI.cow, fill=IMI.cow)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,8) +
  labs(x="log-MAC/mL",title="",fille="IMI") +
  theme(legend.position = "none") +
  geom_vline(xintercept = optimal_mac_C3_IMI_cow, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 


# % Neu

IMI.4_cow_C3 <- Pc.by.cow.C3 %>%
  filter(!is.na(PERCENT.NEU.CO.COW) & !is.na(IMI.cow)) %>%
  ggplot(aes(x=PERCENT.NEU.CO.COW, group=IMI.cow, fill=IMI.cow)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,100) +
  labs(x="% NEU",title="",fille="IMI") +
  theme(legend.position = "none") +
  geom_vline(xintercept = optimal_per_neu_C3_IMI_cow, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 


# % Lym

IMI.5_cow_C3 <- Pc.by.cow.C3 %>%
  filter(!is.na(PERCENT.LYM.CO.COW) & !is.na(IMI.cow)) %>%
  ggplot(aes(x=PERCENT.LYM.CO.COW, group=IMI.cow, fill=IMI.cow)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,100) +
  labs(x="% LYM",title="",fille="IMI") +
  theme(legend.position = "none") +
  geom_vline(xintercept = optimal_per_lym_C3_IMI_cow, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 


# % Mac

IMI.6_cow_C3 <- Pc.by.cow.C3 %>%
  filter(!is.na(PERCENT.MAC.CO.COW) & !is.na(IMI.cow)) %>%
  ggplot(aes(x=PERCENT.MAC.CO.COW, group=IMI.cow, fill=IMI.cow)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,100) +
  labs(x="% MAC",title="",fille="IMI") +
  theme(legend.position = "none") +
  geom_vline(xintercept = optimal_per_mac_C3_IMI_cow, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 

# Multipanel figure

IMI_cow_C3 <- ggarrange(IMI.1_cow_C3 + rremove("ylab"), IMI.2_cow_C3 + rremove("ylab"), IMI.3_cow_C3 + rremove("ylab"),
                            IMI.4_cow_C3 + rremove("ylab"), IMI.5_cow_C3 + rremove("ylab"), IMI.6_cow_C3 + rremove("ylab"),
                            ncol = 3, nrow = 2,align = "h",labels = c("A2", "B2", "C2","D2","E2","F2","G2"))

## Warning: Removed 5 rows containing non-finite outside the scale range
## (`stat_density()`).
## Removed 1 row containing non-finite outside the scale range (`stat_density()`).

IMI_cow_C3_annotated <- annotate_figure(IMI_cow_C3,
                                            top = text_grob("Postpartum: 3 DIM", color = "black",face="bold",size=20,
                                                            hjust = 0, x = 0))

IMI_cow_C3_annotated

ggsave(plot = last_plot(),"./figures/DSCC.IMI_cow_C3.png",width = 20, height = 10, units = "cm") 


# ---- c5 ----

# Neu

IMI.1_cow_C5 <- Pc.by.cow.C5 %>%
  filter(!is.na(log.TOTAL.NEU.CO.COW) & !is.na(IMI.cow)) %>%
  ggplot(aes(x=log.TOTAL.NEU.CO.COW, group=IMI.cow, fill=IMI.cow)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,8) +
  labs(x="log-NEU/mL",title="",fille="IMI") +
  theme(legend.position = "none") +
  geom_vline(xintercept = optimal_neu_C5_IMI_cow, linetype="dotted", size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 

# Lym

IMI.2_cow_C5 <-Pc.by.cow.C5 %>%
  filter(!is.na(log.TOTAL.LYM.CO.COW) & !is.na(IMI.cow)) %>%
  ggplot(aes(x=log.TOTAL.LYM.CO.COW, group=IMI.cow, fill=IMI.cow)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,8) +
  labs(x="log-LYM/mL",title="",fille="IMI") +
  theme(legend.position = "none") +
  geom_vline(xintercept = optimal_lym_C5_IMI_cow, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 


# Mac 

IMI.3_cow_C5 <- Pc.by.cow.C5 %>%
  filter(!is.na(log.TOTAL.MAC.CO.COW) & !is.na(IMI.cow)) %>%
  ggplot(aes(x=log.TOTAL.MAC.CO.COW, group=IMI.cow, fill=IMI.cow)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,8) +
  labs(x="log-MAC/mL",title="",fille="IMI") +
  theme(legend.position = "none") +
  geom_vline(xintercept = optimal_mac_C5_IMI_cow, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 


# % Neu

IMI.4_cow_C5 <- Pc.by.cow.C5 %>%
  filter(!is.na(PERCENT.NEU.CO.COW) & !is.na(IMI.cow)) %>%
  ggplot(aes(x=PERCENT.NEU.CO.COW, group=IMI.cow, fill=IMI.cow)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,100) +
  labs(x="% NEU",title="",fille="IMI") +
  theme(legend.position = "none") +
  geom_vline(xintercept = optimal_per_neu_C5_IMI_cow, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 


# % Lym

IMI.5_cow_C5 <- Pc.by.cow.C5 %>%
  filter(!is.na(PERCENT.LYM.CO.COW) & !is.na(IMI.cow)) %>%
  ggplot(aes(x=PERCENT.LYM.CO.COW, group=IMI.cow, fill=IMI.cow)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,100) +
  labs(x="% LYM",title="",fille="IMI") +
  theme(legend.position = "none") +
  geom_vline(xintercept = optimal_per_lym_C5_IMI_cow, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 


# % Mac

IMI.6_cow_C5 <- Pc.by.cow.C5 %>%
  filter(!is.na(PERCENT.MAC.CO.COW) & !is.na(IMI.cow)) %>%
  ggplot(aes(x=PERCENT.MAC.CO.COW, group=IMI.cow, fill=IMI.cow)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,100) +
  labs(x="% MAC",title="",fille="IMI") +
  theme(legend.position = "none") +
  geom_vline(xintercept = optimal_per_mac_C5_IMI_cow, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 

# Multipanel figure

IMI_cow_C5 <- ggarrange(IMI.1_cow_C5 + rremove("ylab"), IMI.2_cow_C5 + rremove("ylab"), IMI.3_cow_C5 + rremove("ylab"),
                            IMI.4_cow_C5 + rremove("ylab"), IMI.5_cow_C5 + rremove("ylab"), IMI.6_cow_C5 + rremove("ylab"),
                            ncol = 3, nrow = 2,align = "h",labels = c("A3", "B3", "C3","D3","E3","F3","G3"))

## Warning: Removed 6 rows containing non-finite outside the scale range
## (`stat_density()`).

IMI_cow_C5_annotated <- annotate_figure(IMI_cow_C5,
                                            top = text_grob("Postpartum: 5 DIM", color = "black",face="bold",size=20,
                                                            hjust = 0, x = 0))

IMI_cow_C5_annotated

ggsave(plot = last_plot(),"./figures/DSCC.IMI_cow_C5.png",width = 20, height = 10, units = "cm") 


DSCC.IMI_all <- ggarrange(IMI_cow_doff_annotated + rremove("ylab"), IMI_cow_C3_annotated + rremove("ylab"), IMI_cow_C5_annotated + rremove("ylab"),
                          ncol = 1, nrow = 3, common.legend = TRUE,align = "h",widths = 4)


ggsave(plot = last_plot(),"./figures/DSCC.IMI_cow_all.png",width = 20, height = 30, units = "cm")


# ---- Subclinical mastitis cow-level -----

# Neu

SCM.1_cow_doff <- Doff.by.cow.3 %>%
  filter(!is.na(log.TOTAL.NEU.CO.COW) & !is.na(delaval.SCM.CO.COW)) %>%
  ggplot(aes(x=log.TOTAL.NEU.CO.COW, group=delaval.SCM.CO.COW, fill=delaval.SCM.CO.COW)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,8) +
  labs(x="log-NEU/mL",title="",fill="High SCC") +
  geom_vline(xintercept = optimal_neu_doff_SCM_cow, linetype="dotted", size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 

# Lym

SCM.2_cow_doff <-Doff.by.cow.3 %>%
  filter(!is.na(log.TOTAL.LYM.CO.COW) & !is.na(delaval.SCM.CO.COW)) %>%
  ggplot(aes(x=log.TOTAL.LYM.CO.COW, group=delaval.SCM.CO.COW, fill=delaval.SCM.CO.COW)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,8) +
  labs(x="log-LYM/mL",title="",fill="High SCC") +
  geom_vline(xintercept = optimal_lym_doff_SCM_cow, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 


# Mac 

SCM.3_cow_doff <- Doff.by.cow.3 %>%
  filter(!is.na(log.TOTAL.MAC.CO.COW) & !is.na(delaval.SCM.CO.COW)) %>%
  ggplot(aes(x=log.TOTAL.MAC.CO.COW, group=delaval.SCM.CO.COW, fill=delaval.SCM.CO.COW)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,8) +
  labs(x="log-MAC/mL",title="",fill="High SCC") +
  geom_vline(xintercept = optimal_mac_doff_SCM_cow, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 


# % Neu

SCM.4_cow_doff <- Doff.by.cow.3 %>%
  filter(!is.na(PERCENT.NEU.CO.COW) & !is.na(delaval.SCM.CO.COW)) %>%
  ggplot(aes(x=PERCENT.NEU.CO.COW, group=delaval.SCM.CO.COW, fill=delaval.SCM.CO.COW)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,100) +
  labs(x="% NEU",title="",fill="High SCC") +
  geom_vline(xintercept = optimal_per_neu_doff_SCM_cow, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 


# % Lym

SCM.5_cow_doff <- Doff.by.cow.3 %>%
  filter(!is.na(PERCENT.LYM.CO.COW) & !is.na(delaval.SCM.CO.COW)) %>%
  ggplot(aes(x=PERCENT.LYM.CO.COW, group=delaval.SCM.CO.COW, fill=delaval.SCM.CO.COW)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,100) +
  labs(x="% LYM",title="",fill="High SCC") +
  geom_vline(xintercept = optimal_per_lym_doff_SCM_cow, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 


# % Mac

SCM.6_cow_doff <- Doff.by.cow.3 %>%
  filter(!is.na(PERCENT.MAC.CO.COW) & !is.na(delaval.SCM.CO.COW)) %>%
  ggplot(aes(x=PERCENT.MAC.CO.COW, group=delaval.SCM.CO.COW, fill=delaval.SCM.CO.COW)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,100) +
  labs(x="% MAC",title="",fill="High SCC") +
  geom_vline(xintercept = optimal_per_mac_doff_SCM_cow, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 

# Multipanel figure

SCM_cow_doff <- ggarrange(SCM.1_cow_doff + rremove("ylab"), SCM.2_cow_doff + rremove("ylab"), SCM.3_cow_doff + rremove("ylab"),
                                      SCM.4_cow_doff + rremove("ylab"), SCM.5_cow_doff + rremove("ylab"), SCM.6_cow_doff + rremove("ylab"),
                                      ncol = 3, nrow = 2, common.legend = TRUE,align = "h",labels = c("A1", "B1", "C1","D1","E1","F1","G1"))

## Warning: Removed 5 rows containing non-finite outside the scale range
## (`stat_density()`).

## Warning: Removed 5 rows containing non-finite outside the scale range
## (`stat_density()`).

## Warning: Removed 1 row containing non-finite outside the scale range
## (`stat_density()`).

SCM_cow_doff_annotated <- annotate_figure(SCM_cow_doff,
                                                      top = text_grob("Dry-off", color = "black",face="bold",size=20,
                                                                      hjust = 0, x = 0,y=-0.75))

SCM_cow_doff_annotated

ggsave(plot = last_plot(),"./figures/DSCC.SCM_cow_doff.png",width = 20, height = 10, units = "cm") 


# ---- C3 -----

# Neu

SCM.1_cow_C3 <- Pc.by.cow.C3 %>%
  filter(!is.na(log.TOTAL.NEU.CO.COW) & !is.na(delaval.SCM.CO.COW)) %>%
  ggplot(aes(x=log.TOTAL.NEU.CO.COW, group=delaval.SCM.CO.COW, fill=delaval.SCM.CO.COW)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,8) +
  labs(x="log-NEU/mL",title="",fill="High SCC") +
  theme(legend.position = "none") +
  geom_vline(xintercept = optimal_neu_C3_SCM_cow, linetype="dotted", size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 

# Lym

SCM.2_cow_C3 <-Pc.by.cow.C3 %>%
  filter(!is.na(log.TOTAL.LYM.CO.COW) & !is.na(delaval.SCM.CO.COW)) %>%
  ggplot(aes(x=log.TOTAL.LYM.CO.COW, group=delaval.SCM.CO.COW, fill=delaval.SCM.CO.COW)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,8) +
  labs(x="log-LYM/mL",title="",fill="High SCC") +
  theme(legend.position = "none") +
  geom_vline(xintercept = optimal_lym_C3_SCM_cow, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 


# Mac 

SCM.3_cow_C3 <- Pc.by.cow.C3 %>%
  filter(!is.na(log.TOTAL.MAC.CO.COW) & !is.na(delaval.SCM.CO.COW)) %>%
  ggplot(aes(x=log.TOTAL.MAC.CO.COW, group=delaval.SCM.CO.COW, fill=delaval.SCM.CO.COW)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,8) +
  labs(x="log-MAC/mL",title="",fill="High SCC") +
  theme(legend.position = "none") +
  geom_vline(xintercept = optimal_mac_C3_SCM_cow, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 


# % Neu

SCM.4_cow_C3 <- Pc.by.cow.C3 %>%
  filter(!is.na(PERCENT.NEU.CO.COW) & !is.na(delaval.SCM.CO.COW)) %>%
  ggplot(aes(x=PERCENT.NEU.CO.COW, group=delaval.SCM.CO.COW, fill=delaval.SCM.CO.COW)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,100) +
  labs(x="% NEU",title="",fill="High SCC") +
  theme(legend.position = "none") +
  geom_vline(xintercept = optimal_per_neu_C3_SCM_cow, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 


# % Lym

SCM.5_cow_C3 <- Pc.by.cow.C3 %>%
  filter(!is.na(PERCENT.LYM.CO.COW) & !is.na(delaval.SCM.CO.COW)) %>%
  ggplot(aes(x=PERCENT.LYM.CO.COW, group=delaval.SCM.CO.COW, fill=delaval.SCM.CO.COW)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,100) +
  labs(x="% LYM",title="",fill="High SCC") +
  theme(legend.position = "none") +
  geom_vline(xintercept = optimal_per_lym_C3_SCM_cow, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 


# % Mac

SCM.6_cow_C3 <- Pc.by.cow.C3 %>%
  filter(!is.na(PERCENT.MAC.CO.COW) & !is.na(delaval.SCM.CO.COW)) %>%
  ggplot(aes(x=PERCENT.MAC.CO.COW, group=delaval.SCM.CO.COW, fill=delaval.SCM.CO.COW)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,100) +
  labs(x="% MAC",title="",fill="High SCC") +
  theme(legend.position = "none") +
  geom_vline(xintercept = optimal_per_mac_C3_SCM_cow, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 

# Multipanel figure

SCM_cow_C3 <- ggarrange(SCM.1_cow_C3 + rremove("ylab"), SCM.2_cow_C3 + rremove("ylab"), SCM.3_cow_C3 + rremove("ylab"),
                                    SCM.4_cow_C3 + rremove("ylab"), SCM.5_cow_C3 + rremove("ylab"), SCM.6_cow_C3 + rremove("ylab"),
                                    ncol = 3, nrow = 2,align = "h",labels = c("A2", "B2", "C2","D2","E2","F2","G2"))

## Warning: Removed 5 rows containing non-finite outside the scale range
## (`stat_density()`).
## Removed 1 row containing non-finite outside the scale range (`stat_density()`).

SCM_cow_C3_annotated <- annotate_figure(SCM_cow_C3,
                                                    top = text_grob("Postpartum: 3 DIM", color = "black",face="bold",size=20,
                                                                    hjust = 0, x = 0))

SCM_cow_C3_annotated

ggsave(plot = last_plot(),"./figures/DSCC.SCM_cow_C3.png",width = 20, height = 10, units = "cm") 


# ---- c5 ----

# Neu

SCM.1_cow_C5 <- Pc.by.cow.C5 %>%
  filter(!is.na(log.TOTAL.NEU.CO.COW) & !is.na(delaval.SCM.CO.COW)) %>%
  ggplot(aes(x=log.TOTAL.NEU.CO.COW, group=delaval.SCM.CO.COW, fill=delaval.SCM.CO.COW)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,8) +
  labs(x="log-NEU/mL",title="",fill="High SCC") +
  theme(legend.position = "none") +
  geom_vline(xintercept = optimal_neu_C5_SCM_cow, linetype="dotted", size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 

# Lym

SCM.2_cow_C5 <-Pc.by.cow.C5 %>%
  filter(!is.na(log.TOTAL.LYM.CO.COW) & !is.na(delaval.SCM.CO.COW)) %>%
  ggplot(aes(x=log.TOTAL.LYM.CO.COW, group=delaval.SCM.CO.COW, fill=delaval.SCM.CO.COW)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,8) +
  labs(x="log-LYM/mL",title="",fill="High SCC") +
  theme(legend.position = "none") +
  geom_vline(xintercept = optimal_lym_C5_SCM_cow, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 


# Mac 

SCM.3_cow_C5 <- Pc.by.cow.C5 %>%
  filter(!is.na(log.TOTAL.MAC.CO.COW) & !is.na(delaval.SCM.CO.COW)) %>%
  ggplot(aes(x=log.TOTAL.MAC.CO.COW, group=delaval.SCM.CO.COW, fill=delaval.SCM.CO.COW)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,8) +
  labs(x="log-MAC/mL",title="",fill="High SCC") +
  theme(legend.position = "none") +
  geom_vline(xintercept = optimal_mac_C5_SCM_cow, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 


# % Neu

SCM.4_cow_C5 <- Pc.by.cow.C5 %>%
  filter(!is.na(PERCENT.NEU.CO.COW) & !is.na(delaval.SCM.CO.COW)) %>%
  ggplot(aes(x=PERCENT.NEU.CO.COW, group=delaval.SCM.CO.COW, fill=delaval.SCM.CO.COW)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,100) +
  labs(x="% NEU",title="",fill="High SCC") +
  theme(legend.position = "none") +
  geom_vline(xintercept = optimal_per_neu_C5_SCM_cow, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 


# % Lym

SCM.5_cow_C5 <- Pc.by.cow.C5 %>%
  filter(!is.na(PERCENT.LYM.CO.COW) & !is.na(delaval.SCM.CO.COW)) %>%
  ggplot(aes(x=PERCENT.LYM.CO.COW, group=delaval.SCM.CO.COW, fill=delaval.SCM.CO.COW)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,100) +
  labs(x="% LYM",title="",fill="High SCC") +
  theme(legend.position = "none") +
  geom_vline(xintercept = optimal_per_lym_C5_SCM_cow, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 


# % Mac

SCM.6_cow_C5 <- Pc.by.cow.C5 %>%
  filter(!is.na(PERCENT.MAC.CO.COW) & !is.na(delaval.SCM.CO.COW)) %>%
  ggplot(aes(x=PERCENT.MAC.CO.COW, group=delaval.SCM.CO.COW, fill=delaval.SCM.CO.COW)) +
  geom_density(adjust=1.5, alpha=.4) +
  xlim(0,100) +
  labs(x="% MAC",title="",fill="High SCC") +
  theme(legend.position = "none") +
  geom_vline(xintercept = optimal_per_mac_C5_SCM_cow, linetype="dotted", 
             size=1.5) +
  scale_fill_manual(values = c("yellow","red"),labels=c('No', 'Yes')) 

# Multipanel figure

SCM_cow_C5 <- ggarrange(SCM.1_cow_C5 + rremove("ylab"), SCM.2_cow_C5 + rremove("ylab"), SCM.3_cow_C5 + rremove("ylab"),
                                    SCM.4_cow_C5 + rremove("ylab"), SCM.5_cow_C5 + rremove("ylab"), SCM.6_cow_C5 + rremove("ylab"),
                                    ncol = 3, nrow = 2,align = "h",labels = c("A3", "B3", "C3","D3","E3","F3","G3"))

## Warning: Removed 6 rows containing non-finite outside the scale range
## (`stat_density()`).

SCM_cow_C5_annotated <- annotate_figure(SCM_cow_C5,
                                                    top = text_grob("Postpartum: 5 DIM", color = "black",face="bold",size=20,
                                                                    hjust = 0, x = 0))


ggsave(plot = last_plot(),"./figures/DSCC.SCM_cow_C5.png",width = 20, height = 10, units = "cm") 



DSCC.SCM_all <- ggarrange(SCM_cow_doff_annotated + rremove("ylab"), SCM_cow_C3_annotated + rremove("ylab"), SCM_cow_C5_annotated + rremove("ylab"),
                          ncol = 1, nrow = 3, common.legend = TRUE,align = "h",widths = 4)

DSCC.SCM_all

ggsave(plot = last_plot(),"./figures/DSCC.SCM_cow_all.png",width = 20, height = 30, units = "cm")